- Email: [email protected]
Electrocatalytic activity of perovskite-based cathodes for solid oxide fuel cells
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Electrocatalytic activity of perovskite-based cathodes for solid oxide fuel cells D. Clematis a,*, A. Barbucci a,b, S. Presto b, M. Viviani b, M.P. Carpanese a,b di Genova, Via all’Opera Pia 15, Department of Civil, Chemical and Environmental Engineering (DICCA), Universita 16145 Genova, Italy b Institute of Condensed Matter Chemistry and Technology for Energy (ICMATE), National Research Council (CNR), c/o DICCA-UNIGE, Via all’Opera Pia 15, 16145 Genova, Italy a
article info
abstract
Article history:
A very active cathode material for intermediate temperature - solid oxide fuel cells (IT-
Received 12 October 2018
SOFCs) is obtained by mixing La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF) and Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF)
Received in revised form
powders. Three different volume ratios are considered: BSCF-LSCF 50-50 v/v% (BL50), BSCF-
7 January 2019
LSCF 70-30 v/v% (BL70) and 30e70 v/v% (BL30).
Accepted 10 January 2019 Available online 7 February 2019
The electrodes are slurry coated on Ce0.8Sm0.2O2-d electrolyte and sintered at 1100 C. After the sintering step XRD-analyses highlight relevant cation inter-diffusion within the mixed powders. As a result, an enhanced activity of BL30-BL70 electrodes towards oxygen
Keywords:
reduction reaction is detected in comparison to LSCF or BSCF pure powders. A polarization
LSCF
resistance of 0.021 U cm2 at 650 C for BL70 is obtained, one of the lowest value reported in
BSCF
literature for SOFC cathodes. Furthermore, all the electrodes show lower activation energy
Cathode
than the two reference materials in the considered temperature range (500e650 C) and
Distribution of relaxation times
two different kinetic regimes are identified at the extremes of this range. Effect of the
Overpotential
applied overpotential (0e0.3 V) on the electrode kinetic is also investigated.
Performance degradation
After a preliminary ageing, performed at 650 C for 200 h by applying a current density of 200 mA cm-2, the electrodes preserve a remarkable performance as IT-SOFC cathodes, despite an initial degradation. A stable value of 0.048 U cm2 of polarization resistance for the sample richer in BSCF is recorded. © 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction High electrocatalytic activity for oxygen reduction reaction (ORR) and slow degradation rate are two extremely important requirements in cathodes for intermediate temperature - solid oxide fuel cells (IT-SOFC) [1]. The main question, still unanswered, is whether it is possible to meet both of them in the same system. Many studies have been dedicated to the development of new materials, and in particular perovskites
are attractive because of their high ionic and electronic conductivity, as well as for their high oxygen surface exchange coefficient [1e11]. La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF) and Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF) are among the most studied perovskites for electrochemical application. In detail, LSCF is attractive for its high electronic and good ionic conductivities [12]. However, at the operating temperature of zirconia-based SOFC, cation inter-diffusion (La, Sr) causes the formation of insulating phases such as La2Zr2O7 or SrZrO3, which negatively affect performances [13e18].
* Corresponding author. Via all’Opera Pia 15, 16145 Genova, Italy. E-mail address: [email protected] (D. Clematis). https://doi.org/10.1016/j.ijhydene.2019.01.128 0360-3199/© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
On the other hand, BSCF, which has a lower electronic conductivity if compared to LSCF, is very active toward ORR owing to its high oxygen vacancies concentration and oxygen diffusion rate, larger than 2 to 200 times the values of other perovskites [19,20]. However, BSCF is penalised by structural instability below 900 C, with formation of different phases with cubic or hexagonal symmetry and with lower activity for oxygen reduction [21,22]. Besides, both materials are affected by Sr or Ba carbonate formation in the presence of CO2 [23], resulting in electrode deactivation [24,25]. Reaction with CO2 reduces the ability of BSCF for oxygen transport [24] and is attributed to the presence of Ba bonds [25]. The approaches applied to reduce cation diffusion and/or phase transition of LSCF and BSCF are different, among these it is noteworthy to recall: (i) surface modification by means of infiltration by a second phase [26e28], (ii) introduction of Asite deficiency [12,29,30] (iii) in situ polarization-exsolution treatment [31], (iv), use of composite electrodes to stabilise different phases [32,33]. While these methods have general applications a further strategy previously applied by some of the authors [34] and rarely reported in literature [35] is pursued in this paper. It consists in the mixing of LSCF and BSCF in order to take advantage of their strengths and reciprocally compensate their aforementioned drawbacks. Solid state inter-diffusion of elements during electrode sintering and even at operating temperature is expected to take place with possible formation of new phases. However, partial substitution of Barium with Lanthanum in BSCF has been reported as a reliable way to stabilise the cubic structure [36]. Moreover, the mentioned preliminary investigation [34] has shown promising results. For these reasons, the system deserves a further investigation in order to have confirmation of the obtained results and to obtain new data on different LSCF/BSCF ratios. It should be highlighted that, once proven high and stable activity toward ORR, further enhancement of the sought electrochemical properties might be obtained by applying some of the listed general methods, for instance infiltration. Then, this study deals with structure, performance and stability of LSCF-BSCF-based cathodes. The structure and morphology of the electrodes were investigated by X-ray diffraction (XRD) and Scanning Electronic Microscopy (SEM). Electrochemical impedance spectroscopy (EIS) measurements at different temperatures and cathodic overpotentials were carried out to study the electrode performance and ORR kinetic mechanism. Ageing tests in simulated cathodic condition of operation were also performed to compare stability over time.
Experimental Mixture of Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF, Treibacher Industrie AG) and La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF, Fuel Cell Materials) commercial powders with three different volume ratios 50-50 v/v% (BL50), 70-30 v/v% (BL70) and 30e70 v/v% (BL30), were prepared through low energy ball milling in distilled water for 20 h at room temperature. Starting powders were weighted and their skeletal volumes were calculated by considering their densities, i.e. 6.01 g cm3 for LSCF (Pearson's Crystal
6213
Database e 1822311) and 5.19 g cm3 for BSCF (Pearson's Crystal Database e 1816873). For each mixture (15 g in a 100 ml PE bottle), 5 g of distilled water was added and zirconia beads of three different diameters were used for ball milling: f ¼ 15 mm (active material/zirconia mass ratio 1:2), f ¼ 10 mm (active material/zirconia mass ratio 1:2.4) and f ¼ 5 mm (active material/zirconia mass ratio 1:2.6). After milling, the powders were freeze dried for 24 h at 50 C and 0.01 mbar (Labconco FreeZone 2.5) and sieved for 30 min to be separated from beads. Ce0.8Sm0.2O2-d powder (SDC20 e HP Fuel Cell Materials) was used for electrolyte preparation; it was cold pressed at 60 MPa, sintered in one step at 1450 C for 5 h in air, resulting in electrolyte disks with a diameter of 25 mm and a thickness of 1.2 mm. An ink for slurry coating of the electrode on the electrolyte disks was prepared by mixing the above powders in a mortar with alpha-therpineol (tech grade 90%, Sigma Aldrich; 50 drops in 15 g of powder). Cells were prepared in the threeelectrode configuration and symmetry between working electrode (WE) and counter electrode (CE) was achieved by using a home-made mask system. The reference electrode (RE) was placed on the WE side of the cell at a distance of 3 times the electrolyte thickness, as suggested in the literature [37,38] to avoid polarization problems during impedance measurements. All cathodes were sintered at 1100 C for 2 h, obtaining a geometric area of 0.28 cm2 and a thickness between 50 and 70 mm after sintering. In order to identify the crystalline phases and check the cathodes stability during operation, electrodes were analysed by X-ray diffraction (XRD, Cubix, Panalytical), using Cu-Ka incident radiation. Patterns were collected over the angular range 10 < 2q < 70 , in a constant scan mode with steps of 0.02 and a counting time of 7 s. XRD analysis was carried out on cathodes deposited on SDC electrolyte (half cells). Polished cross sections of same samples were also analysed by Scanning Electron Microscopy (SEM, 1450-vp, LEO) and by an energy-dispersive electron microprobe (EDX, INCA 300, Oxford Instruments). Electrochemical impedance spectroscopy (EIS) measurements were carried out in a three-electrode configuration to investigate the cathode catalytic activity; to this purpose, the cells were placed in a lab-manufactured test station [39,40]. A Pt mesh was placed on the surface of the WE and CE as current collector and Pt wires provided the connection to a potentiostat coupled to a frequency response analyser (Autolab PSSTAT302N). Impedance data were collected in the temperature range of 500e650 C at open circuit voltage (OCV) and at applied cathodic overpotentials (up to 0.3 V), with a perturbation of 10 mV in the 0.1e100 kHz frequency range. Ageing tests were performed at 650 C by applying a cathodic direct current (dc) of 200 mA cm2 and checking the system evolution by impedance measurements at OCV, after interruption of the dc load at different times. The cathode behavior was investigated through the analysis of the impedance spectra following two different approaches: equivalent circuits (ZView®, Scribner Associates Inc.) and distribution of relaxation time (DRT). DRT analysis is a deconvolution techniques, which allows one to point out the larger contribution to the frequency
6214
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
dispersion [41,42]. The distribution can be obtained by solving an equation of this type: Z∞ Zðui Þ ¼ R∞ þ Rp ∞
GðtÞ dlnt 1 þ jui
(1)
where Zðui Þ are the experimental data, R∞ is the resistance at high frequency, RP the polarization resistance and G(t) is a normalised function: Z∞ GðtÞdlnt ¼ 1
(2)
∞
Nevertheless this is an ill-posed inverse problem, and many different G(t) solutions are possible or the solution might be unstable, so that different approaches for the regularisation of the solution are available in literature [43e46]. In this paper the DRT technique was applied following Ciucci group's approach and tool [47].
Results and discussion Structural results Fig. 1a shows the XRD profile of BL30, BL50 and BL70 cathodes both as-sintered (AS) and after testing (AT). The profile of a 5050 v/v% mixture of starting powders is also reported for comparison. The structures of starting powders are confirmed as cubic (space group: Pm3m) for BSCF and trigonal (space group: R3c : H) for LSCF. Fig. 1b focuses on one angular range where no peaks from the substrate are visible. In general, both cubic and trigonal structures can be identified in the pattern of all samples, indicated in Fig. 1b as vertical lines. In AS samples, the peaks of the trigonal structure are slightly shifted (<0.08 deg) with respect to the LSCF, while for the cubic phase a significant shift with respect to those of BSCF is observed. The 2q shift is larger in BL30 (y1.2 deg) than in BL50 (y1.15 deg) and in BL70 (y0.8 deg). After testing, BL30 and BL50 are substantially unchanged. On the contrary, in BL70 the peaks of cubic phase are shifted back toward lower 2q angles, while the trigonal phase almost disappeared (see Fig. 1(b3)). These findings can be qualitatively explained by cationic interdiffusion between the two perovskite structures, taking place during sintering at 1100 C. As a result, the stoichiometry of each phase might deviate from the nominal and cannot be determined from standard XRD. Taking into account the incorporation of some smaller La3þ ions (r ¼ 136 pm, CN ¼ 12) replacing larger Sr2þ ions (r ¼ 144 pm, CN ¼ 12) at A-sites of BSCF, a unit cell volume decrease resulting in the positive shift of diffraction peaks position of the cubic structure might be expected. Similarly, the incorporation of Sr2þ replacing La3þ at A-sites of LSCF would cause lattice expansion, in this case balanced by shrinkage due to the formation of oxygen vacancies as charge compensation defect. The overall result is a small shift for the peaks of trigonal structure. By increasing the starting volume fraction of BSCF, the cubic structure prevails with the same interdiffusion and
Fig. 1 e (a) XRD profiles of samples, listed from base to top of diagram: BL70 (AS, dark cyan), BL70 (AT, magenta), BL50 (AS, cyan), BL50 (AT, blue), BL30 (AS, red) BL30 (AT, green), Mixture of starting powders (black). AS ¼ as sintered, AT ¼ after electrochemical test. (b) Magnification. (-) ¼ BSCF, (:) ¼ LSCF, (*) ¼ SDC, (d) ¼ cubic and (d. d) ¼ trigonal. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
incorporation phenomena taking place during sintering. On ageing, the observed backward shift of cubic phase peaks might be explained by progressive incorporation of La, i.e. formation of BLSCF, compensated by elimination of oxygen vacancies, causing lattice expansion. SEM analyses showed that BL30 and BL50 electrode layers were adherent to the electrolyte substrate and with regular shape and thickness, in the range 50÷60 mm. The two phases appeared uniformly mixed although some areas with finer grain size, about 0.4 mm diameter, and others organised in larger clusters, about 2e4 mm diameter, can be observed. Given the particle size and shape of starting
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
powders, i.e. 200 nm spherical for LSCF and about 3 mm faceted for BSCF [34], the two different morphologies observed can be respectively ascribed to LSCF-like or BSCF-like structure. Coherently, the relative amount of larger clusters increased with starting amount of BSCF. These features were not affected by electrochemical testing and ageing and are clearly visible in Fig. 2a and b. It is worth noting that the mean atomic number of La0.6Sr0.4Co0.2Fe0.8O3 and Ba0.5Sr0.5Co0.8Fe0.2O3, calculated by the Bu¨chner formula [48], are very close and therefore will appear with the same grey tone in back scattered electron (BSE) images. Sample BL70 showed good adhesion to the substrate (Fig. 2c), with irregular thickness, ranging between 50 and 70 mm. The microstructure is characterized by the presence of large clusters (>30 mm) and pores with diameter in the 2e10 mm range. After testing the electrode layer was adherent to the electrolyte. Thickness was reduced, between 30 and 40 mm, due to further growth of clusters, with also formation of large voids (Fig. 2d). No secondary phases were detected and the composition obtained by EDX was quite homogeneous within each electrode and corresponding to the average nominal one. However, the formation of small amounts of secondary phases, below the detection limit of the technique and instrumentation used, cannot be ruled out in all samples.
Electrochemical results Electrochemical impedance spectroscopy investigation (EIS) was carried out in three-electrode configuration considering a temperature range between 500 and 650 C (DT ¼ 25 C). Fig. 3a shows the typical impedance spectra for each composition at 650 C in air, compared to reference materials. In these
6215
operating conditions, the best performances were obtained by BL70 and BL50, with a RP equal to 0.021 and 0.027 U cm2, respectively. Nevertheless, also BL30 mixture provided a significant improvement of electrode activity with a RP value equal to 0.045 U cm2; indeed, for starting pure commercial LSCF and BSCF electrodes, measured RP were, 0.46 U cm2 and 0.21 U cm2, respectively. All the RP values obtained for new electrodes are competitive if compared with recent values reported in literature for optimized BSCF-based electrodes [49], which still remains the most active perovskite-based SOFCs cathode [2]. Activation energy (Eact) values calculated from Arrhenius plots are reported in Fig. 3b. In spite of its higher resistivity, BL30 shows the lowest Eact (0.96eV), indicating that this composition becomes more competitive at low temperatures. The sample BL70 has an activation energy value of 1.17 eV close to the pure BSCF (1.18 eV) and lower than pure LSCF (1.52 eV), in agreement with those reported in literature (LSCF 1.45 eV [50], BSCF 1.20 eV [20,51]). The Eact values confirm that all the three new compositions are prone to oxygen reduction reaction, with an enhanced electrocatalytic activity. As reported in Fig. 4a, showing the Nyquist plots of sample BL70 as an example, a remarkable difference in the impedance shape was observed according to the temperature. This strong effect was detected for all the compositions and it indicates that not only polarization resistance, but also the whole kinetic mechanism is very sensitive to this operating parameter, as confirmed by the analysis of distribution of relaxation time (DRT), performed on EIS data at OCV, and illustrated in Fig. 4b. This not trivial behavior prevented the use of a unique and reliable equivalent circuit able to describe all the experimental spectra in the whole temperature range. On the other hand,
Fig. 2 e BSE images on cross section of samples: (a) BL30 as sintered, (b) BL50 after electrochemical testing, (c) BL70 as sintered and (d) BL70 after electrochemical testing.
6216
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Fig. 3 e (a) Impedance spectra at 650 C and OCV for BL30, BL50, BL70. The inset shows impedance spectra for pure BSCF and LSCF. (b) Arrhenius plot for the five cathodes and calculated activation energy: BL30 (0.95 eV), BL50 (1.06 eV), BL70 (1.17 eV), BSCF (1.18 eV) and LSCF (1.52 eV).
Fig. 4 e (a) Impedance spectra and (b) corresponding DRT patterns obtained for sample BL70 at OCV, in the range 500e650 C. The inset in (b) shows a magnification of the DRT curves at 600 and 650 C. distribution of relaxation time (DRT) allowed recognition of the change in predominant phenomena for the two different regimes (500 and 650 C). DRT spectra at 500 C showed a main peak at low frequencies (about 10 Hz, Fig. 4b), followed by a tail at mediumhigh frequencies. This is a typical Gerischer behavior, as reported in different studies [52,53], which implies considering the system as co-controlled by coupled oxygen surface exchange and bulk diffusion, according to the following expression [54]: ZG ¼ Rchem
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ jutchem
(3)
Rchem being the chemical resistance and tchem the characteristic time, strictly related to oxygen surface exchange and transport properties. Positive feedback for the Gerischer hypothesis came from the analysis of impedance data through an R1G equivalent circuit fitting, as illustrated in Fig. 5. Increasing the temperature (typically at 600e650 C) the main peak shifted to a medium-high frequency range (around 140 Hz, Fig. 4b) and its intensity significantly decreased,
Fig. 5 e BL30 and BL70 cathode impedance spectra at 500 C and OCV overlaid on the fitting spectra. R1 ¼ (subtracted) electrolyte resistance; GE1 ¼ Gerischer element (named G in the text).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
indicating a thermal activation of the process and a possible kinetic mechanism modification. Furthermore, at these higher temperatures a second well-defined peak appeared at lower frequency (around 1 Hz, Fig. 4b). Based on the peak frequency and data reported in literature, it is reasonable to consider this process related to surface phenomena, such as oxygen adsorption, or to oxygen diffusion at gas phase [1,55,56]. Going more into detail, at 650 C the total resistance decreases compared to 600 C (Fig. 4a), but the area under low frequency peak (strictly linked to resistance phenomena) rises (inset in Fig. 4b), and its maximum value (PLF) moves to lower frequencies, indicating a process hindered by the temperature increasing. The same temperature effect at low frequency was detected by Amin and Karan [57] for a La0.5Ba0.5CoO3-based cathode; they suggested that this behavior it could be due to an unfavorable oxygen adsorption on the electrode surface, or a partial surface coverage at higher temperature. Nevertheless, the identification of the process in this frequency range is not trivial due to the complexity of the system and the multiplicity of phenomena involved, as also reported by Bidrawn et al. [58], it is extremely difficult to distinguish whether diffusion or oxygen adsorption is the rate-limiting step. Considering the kinetic regime change, further investigation was carried out under cathodic polarization condition in the studied range of temperature. Fig. 6 reports the impedance spectra at different cathodic overpotentials (h) at 500 and 650 C,
6217
with the matched distribution of relaxation times (insets) both for BL30 (Fig. 6a and b) and for BL70 sample (Fig. 6c and d). According to Adler model [54], by manipulating relationship (3), it is possible to obtain the imaginary part of impedance from: vq ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 u ðu$t chem Þ þ 1 1 Rchem t Z''G ¼ pffiffiffi 2 ðu$tchem Þ2 þ 1
(4)
and then extrapolating tchem and Rchem fitting experimental data. Their ratio allows to calculate a chemical capacitance: Cchem ¼ tchem =Rchem
(5)
useful to evaluate the electrode bulk ability to store charge in terms of oxygen vacancies. For both samples, at 500 C, Rchem is a good approximation of global RP for all the applied h, confirming the dominant Gerischer behavior introduced above. More into detail, at 500 C, BL30 exhibits the same behavior observed by the authors for LSCF reference electrode, with a RP reduction as h increases. A possible explanation of this behavior, already proposed for LSCF [26], is the increase on ð$$Þ oxygen vacancy concentration [VO ] under cathodic h and their fundamental contribution to oxygen solid-state diffuð$$Þ sion through the electrode bulk. This [VO ] variation is highlighted also by Cchem response (Fig. 7b) that grows under low h.
Fig. 6 e Impedance spectra and matched distribution of relaxation times at 500 and 600 C for sample BL30 (a and b) and for BL70 (c and d) sample. The mean peak at 500 C appears at 11e12 Hz ((a) and (c)); the mean peak at 650 C appears at about 140 Hz ((b) and (d)).
6218
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Fig. 7 e (a) Rchem, and (b) Cchem as functions of applied overpotential at 500 C, derived from Gerischer fitting, for BL30 and BL70.
Although BL30 shows a smooth Cchem decrease at higher h, its final value globally increases if compared with that calculated in OCV conditions. For BL70 electrode h has an opposite effect (Rp increases with h), similar to BSCF e behavior investigated by some of the authors in previous paper [26]. This work and other papers [34,59,60] highlighted that an applied overpotential can cause ð$$Þ oxygen vacancy cluster formation as [VO ] increases, with consequent losses for oxygen reduction activity of the electrode, explaining why the Cchem, starting increase is not able to lead an improvement in electrode activity. When the system is brought at 650 C the more resistive process was located at medium-high frequency (Fig. 6b and d) pointing out a kinetic change compared to 500 C behavior. Based on DRT analysis and on a previous study [34], the circuit R1 (RHFQHF) (RLFQLF) was used to fit the impedance results. In this equivalent circuit R1 is the electrolyte resistance, RHF and QHF are the resistance and constant phase element (CPE) at high frequency (~140 Hz) and RLF and QLF describe the system at low frequency. According to Ref. [61] a capacitance was
calculated from CPE parameters. All these parameters are reported in Fig. 8a and b, which show that the limiting step at 650 C is the high frequency process (Fig. 6b and d). Also at 650 C, according to their high values, the low frequency capacitance is chemical for both samples, confirming the bulk ability to store charge as oxygen vacancies. This ability is enhanced in composites and increases with BSCF content, as indicated by Cchem values at 650 C and OCV: in fact, Cchem is equal to 0.068 in LSCF, 0.806 in BSCF, raising in BL30, BL50 and BL70 up to 1.14, 5.2 and 12.9 respectively (values expressed in F.cm2). The above-described regime change in the range 500e650 C represents the key to explain the improvement ensured by composite over single-phase cathodes. In fact, remarkable performances of LSCF and BSCF composites are a due to drastic reduction of the polarization resistance associated with oxygen surface exchange and bulk diffusion. This phenomenon can be considered a combined effect of the high oxygen vacancies concentration in BSCF and of the high ionic conductivity in LSCF. The main limitation to conduction at 650 C is still represented by charge transfer at electrodeelectrolyte interface, which could be further improved by electrode fabrication process optimization. After the electrochemical investigation, an ageing test at 650 C under a current load of 200 mA/cm2 was performed to evaluate electrode degradation over the time (Fig. 9). Degradation of electrode performance in the first operating hours is reported as key factor for cathode materials life. Analysis of performance losses in the first 100e300 working hours has been taken as good indication to evaluate electrode stability for longer operating time [62e64]. Consequently, an ageing time of 200 h was chosen to monitor the evolution of the system performing periodic EIS measurements at regular time. All the electrodes showed an initial remarkable degradation, but Rp trend are different for the three compositions. BL30 and BL50 showed a quite linear increase of Rp, while for BL70 after the first 50 h the degradation rate slowed down, remaining the most attractive cathode. The Rp value stabilized close to 0.048 U cm2, that is highly competitive even if compared with resistance data of pure and fresh LSCF and BSCF cathode reported in literature [31,52,55,63,65e71] and
Fig. 8 e Values of resistances (a) and capacitances (b) extracted from the fitting analysis of impedance results obtained at 650 C for BL30 and BL70 samples.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Fig. 9 e Degradation rate for BL30, BL50, BL70 electrodes and BL70 chemical capacitance evolution under an applied current load of 200 mA/cm2 at 650 C.
also with highly-performing BLSCF [63]. Confirmation about stabilization of BL70 is provided by the decrease of degradation rate; indeed, three different intervals were identified between 0 and 24 h, 24e72 h and above 72 h with a slope of 6.67 104, 1.66 104 and 3.58 105 U cm2 h1 respectively. The starting order of magnitude for degradation rate (~104 U cm2 h1) is in accordance with those extrapolated for BSCF and LSCF (103 ÷104 U cm2 h1) by the authors and other papers [26,34,62], while the drop to ~105 U cm2 h1 (R2 ¼ 0.992) represents a stability improvement for electrode systems based on perovskite materials. XRD and SEM investigations did not show structural and morphological modifications during ageing of BL30 and BL50 samples. It is known [72] that, at SOFC operative temperature, segregation of Sr and Co is likely to take place in LSCF. Therefore, the formation of tiny layers of Sr oxides with limited conductivity can be expected at grain boundaries on aging, but cannot be detected due to the resolution of standard SEM and XRD used in this work as reported in the structural analysis section. The increase of Rp (Fig. 9) with aging time in BL30 and BL50 can be ascribed to this phenomenon, which is more pronounced in BL30, i.e. the mixture with highest LSCF content. On the other hand, the observed backward shift of cubic phase XRD peaks during BL70 ageing, explained as progressive incorporation of La compensated by a decrease of oxygen vacancies, find confirmation on the measure of the amount of oxygen vacancies through CChem, which shows a decrease from 10 to 1 F cm2 in the first 50 h followed by a stable value in the remaining hours. Such changes in the bulk of the electrode match well with the BL70 performance losses displayed in the first 50 h of ageing of Fig. 9 and the morphology evolution observed by SEM analysis. As a matter of fact, BL30 and BL50 CChem remain unchanged during ageing, indicating a substantial constant concentration of oxygen vacancies.
Conclusions The effect of composition in performance and stability of La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF) - Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF)
6219
based cathodes was investigated. The mixtures were prepared through ball milling of commercial LSCF and BSCF powders. The cathodes were deposited on Ce0.8Sm0.2O2-d-electrolytes and investigated in SOFC typical conditions of operation (different temperatures and cathodic overpotentials). Microstructural investigation revealed cations interdiffusion between the cubic BSCF and trigonal LSCF structures during sintering, without formation of secondary phases. For BL70, the cubic structure was predominant and showed lattice parameter increase on ageing, compatible with oxygen uptake from atmosphere. Analyses of the impedance results by distribution of relaxation time (DRT) and equivalent circuits showed two kinetic regimes governing the electrocatalytic activity. At low temperature (500 C), the more resistive step was identified at low frequencies and global impedance was fitted through a Gerischer element, pointing out a behavior co-controlled by surface oxygen exchange and bulk diffusion; this model was no longer suitable for the high temperature regime (650 C). The resistance trends for BL30 and BL70 as functions of overpotential appeared different: an increase of h led to a Rchem decrease for BL30 (LSCF-like behavior) and to an increase for BL70 sample (BSCF-like behavior). Like BSCF, also BL70 electrochemical activity decreased under applied overpotential, because of a suppression of vacancy concentration, and this behavior was not mitigated by the presence of a significant (30%) starting amount of LSCF. All cathodes showed very good electrochemical performance. A preliminary ageing test (200 h) was carried out at 650 C; despite the observed degradation for all three investigated cathodes, BL70 Rp showed an almost asymptotic trend, displaying, for the ageing studied time, a final stabilized and low value close to 0.048 U cm2. These first results open the way for a further optimization step to modulate physical and chemical electrode properties, with a more detailed experimental approach based on application of strategies already available to improve electrode stability and activity. Nevertheless, investigations at longer loading time are necessary and planned for the near future, especially for BL70 composition.
references
[1] Adler SB. Factors governing oxygen reduction in solid oxide fuel cell cathodes. Chem Rev 2004;104:4791e844. https:// doi.org/10.1021/cr020724o. [2] Areum J, Junyoung K, Jeeyoung S, Guntae K. Perovskite as a cathode material: a review of its role in solid-oxide fuel cell technology. ChemElectroChem 2016;3:511e30. https:// doi.org/10.1002/celc.201500382. [3] Carpanese MP, Barbucci A, Canu G, Viviani M. BaCe0.85Y0.15O2.925 dense layer by wet powder spraying as electrolyte for SOFC/SOEC applications. Solid State Ionics 2015;269:80e5. https://doi.org/10.1016/j.ssi.2014.11.014. [4] Skinner SJ. Recent advances in Perovskite-type materials for solid oxide fuel cell cathodes. Int J Inorg Mater 2001;3:113e21. https://doi.org/10.1016/S1466-6049(01)00004-6. [5] Viviani M, Canu G, Carpanese MP, Barbucci A, Sanson A, Mercadelli E, et al. Dual cells with mixed protonic-anionic conductivity for reversible SOFC/SOEC operation. Energy
6220
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Procedia 2012;28:182e9. https://doi.org/10.1016/j.egypro. 2012.08.052. Vladikova D, Stoynov Z, Chesnaud A, Thorel A, Viviani M, Barbucci A, et al. Application of yttrium doped barium cerate for improvement of the dual membrane SOFC design. Int J Hydrogen Energy 2014;39:21561e8. https://doi.org/10.1016/j. ijhydene.2014.09.014. da Silva FS, de Souza TM. Novel materials for solid oxide fuel cell technologies: a literature review. Int J Hydrogen Energy 2017;42:26020e36. https://doi.org/10.1016/j.ijhydene. 2017.08.105. Dai Y, Lou Z, Wang Z, Qiao J, Sun W, Sun K. Electrochemical and chemical stability performance improvement of Ba0.5Sr0.5Fe0.91Al0.09O3d cathode for IT-SOFC through the introduction of a GDC interlayer. Int J Hydrogen Energy 2015;40:5939e46. https://doi.org/10.1016/j.ijhydene.2015.02.131. Huang X, Feng J, Abdellatif HRS, Zou J, Zhang G, Ni C. Electrochemical evaluation of double perovskite PrBaCo2xMnxO5þd (x ¼ 0, 0.5, 1) as promising cathodes for IT-SOFCs. Int J Hydrogen Energy 2018;43:8962e71. https://doi.org/10. 1016/j.ijhydene.2018.03.163. Nicollet C, Flura A, Vibhu V, Rougier A, Bassat J-M, Grenier JC. An innovative efficient oxygen electrode for SOFC: Pr6O11 infiltrated into Gd-doped ceria backbone. Int J Hydrogen Energy 2016;41:15538e44. https://doi.org/10.1016/j.ijhydene. 2016.04.024. pez C, Hauch A, Ruiz de Ortiz-Vitoriano N, Bernuy-Lo Larramendi I, Rojo T. Electrochemical characterization of La0.6Ca0.4Fe0.8Ni0.2O3 cathode on Ce0.8Gd0.2O1.9 electrolyte for IT-SOFC. Int J Hydrogen Energy 2014;39:6675e9. https://doi. org/10.1016/j.ijhydene.2014.01.032. Mineshige A, Izutsu J, Nakamura M, Nigaki K, Abe J, Kobune M, et al. Introduction of A-site deficiency into La0.6Sr0.4Co0.2Fe0.8O3ed and its effect on structure and conductivity. Solid State Ionics 2005;176:1145e9. https://doi. org/10.1016/j.ssi.2004.11.021. Kiebach R, Zhang W-W, Zhang W, Chen M, Norrman K, Wang H-J, et al. Stability of La0.6Sr0.4Co0.2Fe0.8O3/Ce0.9Gd0.1O2 cathodes during sintering and solid oxide fuel cell operation. J Power Sources 2015;283:151e61. https://doi.org/10.1016/j. jpowsour.2015.02.064. Wang F, Nishi M, Brito ME, Kishimoto H, Yamaji K, Yokokawa H, et al. Sr and Zr diffusion in LSCF/10GDC/8YSZ triplets for solid oxide fuel cells (SOFCs). J Power Sources 2014;258:281e9. https://doi.org/10.1016/j.jpowsour.2014.02. 046. Wang H, Yakal-Kremski KJ, Yeh T, Rupp GM, Limbeck A, Fleig J, et al. Mechanisms of performance degradation of (La,Sr)(Co,Fe)O3-d solid oxide fuel cell cathodes. J Electrochem Soc 2016;163:F581e5. https://doi.org/10.1149/2.0031607jes. Khan MZ, Mehran MT, Song R-H, Lee J-W, Lee S-B, Lim T-H, et al. Effect of GDC interlayer thickness on durability of solid oxide fuel cell cathode. Ceram Int 2016;42:6978e84. https:// doi.org/10.1016/j.ceramint.2016.01.085. Wang F, Brito ME, Yamaji K, Cho D-H, Nishi M, Kishimoto H, et al. Effect of polarization on Sr and Zr diffusion behavior in LSCF/GDC/YSZ system. Solid State Ionics 2014;262:454e9. https://doi.org/10.1016/j.ssi.2014.04.002. Khan MZ, Mehran MT, Song R-H, Lee S-B, Lim T-H. Effects of applied current density and thermal cycling on the degradation of a solid oxide fuel cell cathode. Int J Hydrogen Energy 2018;43:12346e57. https://doi.org/10.1016/j.ijhydene. 2018.04.175. Araki W, Arai Y, Malzbender J. Transitions of Ba0.5Sr0.5Co0.8Fe0.2O3-d and La0.58Sr0.4Co0.2Fe0.8O3-d. Mater Lett 2014;132:295e7. https://doi.org/10.1016/j.matlet.2014.06.093. Shao Z, Haile SM. A high-performance cathode for the next generation of solid-oxide fuel cells. Nature 2004;431:170.
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
https://doi.org/10.1038/nature02863. https://www.nature. com/articles/nature02863#supplementary-information. Mueller DN, De Souza RA, Weirich TE, Roehrens D, Mayer J, Martin M. A kinetic study of the decomposition of the cubic perovskite-type oxide BaxSr1-xCo0.8Fe0.2O3-d (BSCF) (x ¼ 0.1 and 0.5). Phys Chem Chem Phys 2010;12:10320e8. https:// doi.org/10.1039/c0cp00004c. ng Z, Martynczuk J, Efimov K, Harvey AS, Infortuna A, Ya Kocher P, et al. Oxygen-vacancy-related structural phase transition of Ba0.8Sr0.2Co0.8Fe0.2O3-d. Chem Mater 2011;23:3169e75. https://doi.org/10.1021/cm200373r. Bermudez JM, Garcia-Fayos J, Reina TR, Reed G, Persoon ES, € rtz D, et al. Thermochemical stability of LaxSr1-xCoyFe1Go yO3-d and NiFe2O4-Ce0.8Tb0.2O2-d under real conditions for its application in oxygen transport membranes for oxyfuel combustion. J Membr Sci 2018;562:26e37. https://doi.org/10. 1016/j.memsci.2018.05.012. Arnold M, Wang H, Feldhoff A. Influence of CO2 on the oxygen permeation performance and the microstructure of perovskite-type (Ba0.5Sr0.5)(Co0.8Fe0.2)O3d membranes. J Membr Sci 2007;293:44e52. https://doi.org/10.1016/j.memsci. 2007.01.032. Zhang Y, Yang G, Chen G, Ran R, Zhou W, Shao Z. Evaluation of the CO2 poisoning effect on a highly active cathode SrSc0.175Nb0.025Co0.8O3-d in the oxygen reduction reaction. ACS Appl Mater Interfaces 2016;8:3003e11. https://doi.org/ 10.1021/acsami.5b09780. Giuliano A, Carpanese MP, Clematis D, Boaro M, Pappacena A, Deganello F, et al. Infiltration, overpotential and ageing effects on cathodes for solid oxide fuel cells: La0.6Sr0.4Co0.2Fe0.8O3-d versus Ba0.5Sr0.5Co0.8Fe0.2O3-d. J Electrochem Soc 2017;164:F3114e22. https://doi.org/10.1149/ 2.0161710jes. Ai N, Chen K, Jiang SP. A La0.8Sr0.2MnO3/ La0.6Sr0.4Co0.2Fe0.8O3d coreeshell structured cathode by a rapid sintering process for solid oxide fuel cells. Int J Hydrogen Energy 2017;42:7246e51. https://doi.org/10.1016/j. ijhydene.2016.10.036. Chen J, Wan D, Sun X, Li B, Lu M. Electrochemical impedance spectroscopic characterization of impregnated La0.6Sr0.4Co0.2Fe0.8O3d cathode for intermediate-temperature SOFCs. Int J Hydrogen Energy 2018;43:9770e6. https://doi.org/ 10.1016/j.ijhydene.2018.03.223. Naumovich EN, Zakharchuk K, Obre˛bowski S, Yaremchenko A. (La,Sr)(Fe,Co)O3-based cathode contact materials for intermediate-temperature solid oxide fuel cells. Int J Hydrogen Energy 2017;42:29443e53. https://doi.org/10. 1016/j.ijhydene.2017.09.122. Zhou F, Liu Y, Zhao X, Tang W, Yang S, Zhong S, et al. Effects of cerium doping on the performance of LSCF cathodes for intermediate temperature solid oxide fuel cells. Int J Hydrogen Energy 2018;43:18946e54. https://doi.org/10.1016/j. ijhydene.2018.08.041. Zhang X, Zhang W, Zhang L, Meng J, Meng F, Liu X, et al. Enhanced electrochemical property of La0.6Sr0.4Co0.8Fe0.2O3 as cathode for solid oxide fuel cell by efficient in situ polarization-exsolution treatment. Electrochim Acta 2017;258:1096e105. https://doi.org/10.1016/j.electacta.2017. 11.163. Perry Murray E, Sever MJ, Barnett SA. Electrochemical performance of (La,Sr)(Co,Fe)O3e(Ce,Gd)O3 composite cathodes. Solid State Ionics 2002;148:27e34. https://doi.org/ 10.1016/S0167-2738(02)00102-9. Liu Y, Bi J, Chi B, Pu J, Jian L. Effects of impregnating palladium on catalytic performance of LSCF-GDC composite cathodes for intermediate temperature solid oxide fuel cells. Int J Hydrogen Energy 2016;41:6486e92. https://doi.org/10. 1016/j.ijhydene.2016.02.142.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
[34] Giuliano A, Carpanese MP, Panizza M, Cerisola G, Clematis D, Barbucci A. Characterisation of La0.6Sr0.4Co0.2Fe0.8O3-d e Ba0.5Sr0.5Co0.8Fe0.2O3-d composite as cathode for solid oxide fuel cells. Electrochim Acta 2017;240:258e66. https://doi.org/ 10.1016/j.electacta.2017.04.079. ska E, Tatko M. [35] Mosiałek M, Ke˛dra A, Krzan M, Bielan Ba0.5Sr0.5Co0.8Fe0.2O3edeLa0.6Sr0.4Co0.8Fe0.2O3ed composite cathode for solid oxide fuel cell. Arch Metall Mater 2016;61. https://doi.org/10.1515/amm-2016-0243. [36] Kim J-S, Yeon D-H, Jung DW, Kwak C. A highly active and long-term stable La-doped BaxSr1xCo1yFeyO3d cathode for solid-oxide fuel cells. J Power Sources 2014;249:66e71. https://doi.org/10.1016/j.jpowsour.2013.10.053. [37] Cimenti M, Co AC, Birss VI, Hill JM. Distortions in electrochemical impedance spectroscopy measurements using 3-electrode methods in SOFC. I e Effect of cell geometry. Fuel Cells 2007;7:364e76. https://doi.org/10.1002/ fuce.200700019. [38] Winkler J, Hendriksen PV, Bonanos N, Mogensen M. Geometric requirements of solid electrolyte cells with a reference electrode. J Electrochem Soc 1998;145:1184e92. https://doi.org/10.1149/1.1838436. [39] Carpanese MP, Panizza M, Viviani M, Mercadelli E, Sanson A, Barbucci A. Study of reversible SOFC/SOEC based on a mixed anionic-protonic conductor. J Appl Electrochem 2015;45:657e65. https://doi.org/10.1007/s10800-015-0838-8. [40] Presto S, Barbucci A, Carpanese MP, Viviani M, Marazza R. Electrochemical performance of Ni-based anodes for solid oxide fuel cells. J Appl Electrochem 2009;39:2257e64. https:// doi.org/10.1007/s10800-009-9857-7. [41] Boukamp BA, Rolle A. Analysis and application of distribution of relaxation times in solid state ionics. Solid State Ionics 2017;302:12e8. https://doi.org/10.1016/j.ssi.2016. 10.009. e E. [42] Schichlein H, Mu¨ller AC, Voigts M, Kru¨gel A, Ivers-Tiffe Deconvolution of electrochemical impedance spectra for the identification of electrode reaction mechanisms in solid oxide fuel cells. J Appl Electrochem 2002;32:875e82. https:// doi.org/10.1023/a:1020599525160. € rlin T. Deconvolution and maximum entropy in [43] Ho impedance spectroscopy of noninductive systems. Solid State Ionics 1998;107:241e53. https://doi.org/10.1016/S01672738(98)00008-3. [44] Renaut RA, Baker R, Horst M, Johnson C, Nasir D. Stability and error analysis of the polarization estimation inverse problem for microbial fuel cells. Inverse Probl 2013;29:045006. € nleber M, Weber A, Ivers-Tiffe e E. [45] Schmidt JP, Berg P, Scho The distribution of relaxation times as basis for generalized time-domain models for Li-ion batteries. J Power Sources 2013;221:70e7. https://doi.org/10.1016/j.jpowsour.2012.07. 100. [46] Tuncer E, Macdonald JR. Comparison of methods for estimating continuous distributions of relaxation times. J Appl Phys 2006;99:074106. https://doi.org/10.1063/1.2188053. [47] Wan TH, Saccoccio M, Chen C, Ciucci F. Influence of the discretization methods on the distribution of relaxation times deconvolution: implementing radial basis functions with DRTtools. Electrochim Acta 2015;184:483e99. https:// doi.org/10.1016/j.electacta.2015.09.097. [48] Howell PGT, Davy KMW, Boyde A. Mean atomic number and backscattered electron coefficient calculations for some materials with low mean atomic number. Scanning 1998;20:35e40. https://doi.org/10.1002/sca.1998.4950200105. € rmer H, Meffert M, Sza sz J, Wankmu¨ller F, [49] Almar L, Sto Gerthsen D, et al. Improved phase stability and CO2 poisoning robustness of Y-doped Ba0.5Sr0.5Co0.8Fe0.2O3d SOFC cathodes at intermediate temperatures. ACS Appl
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
6221
Energy Mater 2018;1:1316e27. https://doi.org/10.1021/ acsaem.8b00028. Serra JM, Garcia-Fayos J, Baumann S, Schulze-Ku¨ppers F, Meulenberg WA. Oxygen permeation through tape-cast asymmetric all-La0.6Sr0.4Co0.2Fe0.8O3d membranes. J Membr Sci 2013;447:297e305. https://doi.org/10.1016/j.memsci.2013. 07.030. Duan C, Hook D, Chen Y, Tong J, O'Hayre R. Zr and Y codoped perovskite as a stable, high performance cathode for solid oxide fuel cells operating below 500 C. Energy Environ Sci 2017;10:176e82. https://doi.org/10.1039/c6ee01915c. sz J, Weber A, Ivers-Tiffe e E. Oxygen transport Almar L, Sza kinetics of mixed ionic-electronic conductors by coupling focused ion beam tomography and electrochemical impedance spectroscopy. J Electrochem Soc 2017;164:F289e97. https://doi.org/10.1149/2.0851704jes. Zhang Y, Chen Y, Yan M, Chen F. Reconstruction of relaxation time distribution from linear electrochemical impedance spectroscopy. J Power Sources 2015;283:464e77. https://doi.org/10.1016/j.jpowsour.2015.02.107. Adler SB, Lane JA, Steele BCH. Electrode kinetics of porous mixed-conducting oxygen electrodes. J Electrochem Soc 1996;143:3554e64. https://doi.org/10.1149/1.1837252. s D, Soldati AL, Troiani H, Montenegro-Herna ndez A, Garce Caneiro A, Mogni LV. La/Ba-based cobaltites as IT-SOFC cathodes: a discussion about the effect of crystal structure and microstructure on the O2-reduction reaction. Electrochim Acta 2016;215:637e46. https://doi.org/10.1016/j. electacta.2016.08.132. Im H-N, Choi M-B, Jeon S-Y, Singh B, Song S-J. Oxygen reduction properties of La0.1Sr0.9Co0.8Fe0.2O3-d cathode for SOFC using electrochemical method. Meet Abstr 2014:729. MA2014-01. Amin R, Karan K. Characterization of La0.5Ba0.5CoO3 d as a SOFC cathode material. J Electrochem Soc 2010;157:B285e91. https://doi.org/10.1149/1.3267879. Bidrawn F, Ku¨ngas R, Vohs JM, Gorte RJ. Modeling impedance response of SOFC cathodes prepared by infiltration. J Electrochem Soc 2011;158:B514e25. https://doi.org/10.1149/ 1.3565174. Gangopadhyay S, Masunov AE, Inerbaev T, Mesit J, Guha RK, Sleiti AK, et al. Understanding oxygen vacancy migration and clustering in barium strontium cobalt iron oxide. Solid State Ionics 2010;181:1067e73. https://doi.org/10.1016/j.ssi.2010.05. 014. Tsipis EV, Naumovich EN, Patrakeev MV, Yaremchenko AA, Marozau IP, Kovalevsky AV, et al. Oxygen deficiency, vacancy clustering and ionic transport in (La,Sr)CoO3d. Solid State Ionics 2011;192:42e8. https://doi.org/10.1016/j.ssi.2010.03.029. Hirschorn B, Orazem ME, Tribollet B, Vivier V, Frateur I, Musiani M. Determination of effective capacitance and film thickness from constant-phase-element parameters. Electrochim Acta 2010;55:6218e27. https://doi.org/10.1016/j. electacta.2009.10.065. e E. LongEndler C, Leonide A, Weber A, Tietz F, Ivers-Tiffe Term study of MIEC cathodes for intermediate temperature solid oxide fuel cells. ECS trans 2009;25:2381e90. https:// doi.org/10.1149/1.3205791. Junyoung K, Sihyuk C, Areum J, Young JH, Jeeyoung S, Guntae K. Chemically stable perovskites as cathode materials for solid oxide fuel cells: La-doped Ba0.5Sr0.5Co0.8Fe0.2O3d. ChemSusChem 2014;7:1669e75. https://doi.org/10.1002/cssc.201301401. Wang WG, Mogensen M. High-performance lanthanumferrite-based cathode for SOFC. Solid State Ionics 2005;176:457e62. https://doi.org/10.1016/j.ssi.2004.09.007. Chen C-H, Chang C-L, Hwang B-H. Electrochemical and microstructure characteristics of Ba0.5Sr0.5Co0.8Fe0.2O3d
6222
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
(BSCF) cathodes prepared by citrate precursor method for SOFCs. Mater Chem Phys 2009;115:478e82. https://doi.org/10. 1016/j.matchemphys.2009.01.018. [66] Chen Y, Qian B, Li S, Jiao Y, Tade MO, Shao Z. The influence of impurity ions on the permeation and oxygen reduction properties of Ba0.5Sr0.5Co0.8Fe0.2O3d perovskite. J Membr Sci 2014;449:86e96. https://doi.org/10.1016/j.memsci.2013.08.021. [67] Deganello F, Liotta LF, Marcı` G, Fabbri E, Traversa E. Strontium and iron-doped barium cobaltite prepared by solution combustion synthesis: exploring a mixed-fuel approach for tailored intermediate temperature solid oxide fuel cell cathode materials. Mater Renew Sustain Energy 2013;2:8. https://doi.org/10.1007/s40243-013-0008-z. MO, Shao Z. A top-down strategy [68] Su C, Xu X, Chen Y, Liu Y, Tade for the synthesis of mesoporous Ba0.5Sr0.5Co0.8Fe0.2O3d as a cathode precursor for buffer layer-free deposition on stabilized zirconia electrolyte with a superior electrochemical performance. J Power Sources 2015;274:1024e33. https://doi. org/10.1016/j.jpowsour.2014.10.177.
[69] Sumi H, Ohshiro T, Nakayama M, Suzuki T, Fujishiro Y. Prevention of reaction between (Ba,Sr)(Co,Fe)O3 cathodes and yttria-stabilized zirconica electrolytes for intermediate-temperature solid oxide fuel cells. Electrochim Acta 2015;184:403e9. https://doi.org/10.1016/j.electacta.2015. 10.092. [70] Wang K, Ran R, Zhou W, Gu H, Shao Z, Ahn J. Properties and performance of Ba0.5Sr0.5Co0.8Fe0.2O3dþSm0.2Ce0.8O1.9 composite cathode. J Power Sources 2008;179:60e8. https:// doi.org/10.1016/j.jpowsour.2007.12.051. [71] Zhu X, Xia H, Li Y, Lu¨ Z. A (La,Sr)MnO3 nano-film embedded into (Ba,Sr)(Co,Fe)O3 porous cathode for stability enhancement. Mater Lett 2015;161:549e53. https://doi.org/ 10.1016/j.matlet.2015.09.040. [72] Hardy JS, Coyle CA, Bonnett JF, Templeton JW, Canfield NL, Edwards DJ, et al. Evaluation of cation migration in lanthanum strontium cobalt ferrite solid oxide fuel cell cathodes via in-operando X-ray diffraction. J Mater Chem 2018;6:1787e801. https://doi.org/10.1039/c7ta06856e.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Electrocatalytic activity of perovskite-based cathodes for solid oxide fuel cells D. Clematis a,*, A. Barbucci a,b, S. Presto b, M. Viviani b, M.P. Carpanese a,b di Genova, Via all’Opera Pia 15, Department of Civil, Chemical and Environmental Engineering (DICCA), Universita 16145 Genova, Italy b Institute of Condensed Matter Chemistry and Technology for Energy (ICMATE), National Research Council (CNR), c/o DICCA-UNIGE, Via all’Opera Pia 15, 16145 Genova, Italy a
article info
abstract
Article history:
A very active cathode material for intermediate temperature - solid oxide fuel cells (IT-
Received 12 October 2018
SOFCs) is obtained by mixing La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF) and Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF)
Received in revised form
powders. Three different volume ratios are considered: BSCF-LSCF 50-50 v/v% (BL50), BSCF-
7 January 2019
LSCF 70-30 v/v% (BL70) and 30e70 v/v% (BL30).
Accepted 10 January 2019 Available online 7 February 2019
The electrodes are slurry coated on Ce0.8Sm0.2O2-d electrolyte and sintered at 1100 C. After the sintering step XRD-analyses highlight relevant cation inter-diffusion within the mixed powders. As a result, an enhanced activity of BL30-BL70 electrodes towards oxygen
Keywords:
reduction reaction is detected in comparison to LSCF or BSCF pure powders. A polarization
LSCF
resistance of 0.021 U cm2 at 650 C for BL70 is obtained, one of the lowest value reported in
BSCF
literature for SOFC cathodes. Furthermore, all the electrodes show lower activation energy
Cathode
than the two reference materials in the considered temperature range (500e650 C) and
Distribution of relaxation times
two different kinetic regimes are identified at the extremes of this range. Effect of the
Overpotential
applied overpotential (0e0.3 V) on the electrode kinetic is also investigated.
Performance degradation
After a preliminary ageing, performed at 650 C for 200 h by applying a current density of 200 mA cm-2, the electrodes preserve a remarkable performance as IT-SOFC cathodes, despite an initial degradation. A stable value of 0.048 U cm2 of polarization resistance for the sample richer in BSCF is recorded. © 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction High electrocatalytic activity for oxygen reduction reaction (ORR) and slow degradation rate are two extremely important requirements in cathodes for intermediate temperature - solid oxide fuel cells (IT-SOFC) [1]. The main question, still unanswered, is whether it is possible to meet both of them in the same system. Many studies have been dedicated to the development of new materials, and in particular perovskites
are attractive because of their high ionic and electronic conductivity, as well as for their high oxygen surface exchange coefficient [1e11]. La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF) and Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF) are among the most studied perovskites for electrochemical application. In detail, LSCF is attractive for its high electronic and good ionic conductivities [12]. However, at the operating temperature of zirconia-based SOFC, cation inter-diffusion (La, Sr) causes the formation of insulating phases such as La2Zr2O7 or SrZrO3, which negatively affect performances [13e18].
* Corresponding author. Via all’Opera Pia 15, 16145 Genova, Italy. E-mail address: [email protected] (D. Clematis). https://doi.org/10.1016/j.ijhydene.2019.01.128 0360-3199/© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
On the other hand, BSCF, which has a lower electronic conductivity if compared to LSCF, is very active toward ORR owing to its high oxygen vacancies concentration and oxygen diffusion rate, larger than 2 to 200 times the values of other perovskites [19,20]. However, BSCF is penalised by structural instability below 900 C, with formation of different phases with cubic or hexagonal symmetry and with lower activity for oxygen reduction [21,22]. Besides, both materials are affected by Sr or Ba carbonate formation in the presence of CO2 [23], resulting in electrode deactivation [24,25]. Reaction with CO2 reduces the ability of BSCF for oxygen transport [24] and is attributed to the presence of Ba bonds [25]. The approaches applied to reduce cation diffusion and/or phase transition of LSCF and BSCF are different, among these it is noteworthy to recall: (i) surface modification by means of infiltration by a second phase [26e28], (ii) introduction of Asite deficiency [12,29,30] (iii) in situ polarization-exsolution treatment [31], (iv), use of composite electrodes to stabilise different phases [32,33]. While these methods have general applications a further strategy previously applied by some of the authors [34] and rarely reported in literature [35] is pursued in this paper. It consists in the mixing of LSCF and BSCF in order to take advantage of their strengths and reciprocally compensate their aforementioned drawbacks. Solid state inter-diffusion of elements during electrode sintering and even at operating temperature is expected to take place with possible formation of new phases. However, partial substitution of Barium with Lanthanum in BSCF has been reported as a reliable way to stabilise the cubic structure [36]. Moreover, the mentioned preliminary investigation [34] has shown promising results. For these reasons, the system deserves a further investigation in order to have confirmation of the obtained results and to obtain new data on different LSCF/BSCF ratios. It should be highlighted that, once proven high and stable activity toward ORR, further enhancement of the sought electrochemical properties might be obtained by applying some of the listed general methods, for instance infiltration. Then, this study deals with structure, performance and stability of LSCF-BSCF-based cathodes. The structure and morphology of the electrodes were investigated by X-ray diffraction (XRD) and Scanning Electronic Microscopy (SEM). Electrochemical impedance spectroscopy (EIS) measurements at different temperatures and cathodic overpotentials were carried out to study the electrode performance and ORR kinetic mechanism. Ageing tests in simulated cathodic condition of operation were also performed to compare stability over time.
Experimental Mixture of Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF, Treibacher Industrie AG) and La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF, Fuel Cell Materials) commercial powders with three different volume ratios 50-50 v/v% (BL50), 70-30 v/v% (BL70) and 30e70 v/v% (BL30), were prepared through low energy ball milling in distilled water for 20 h at room temperature. Starting powders were weighted and their skeletal volumes were calculated by considering their densities, i.e. 6.01 g cm3 for LSCF (Pearson's Crystal
6213
Database e 1822311) and 5.19 g cm3 for BSCF (Pearson's Crystal Database e 1816873). For each mixture (15 g in a 100 ml PE bottle), 5 g of distilled water was added and zirconia beads of three different diameters were used for ball milling: f ¼ 15 mm (active material/zirconia mass ratio 1:2), f ¼ 10 mm (active material/zirconia mass ratio 1:2.4) and f ¼ 5 mm (active material/zirconia mass ratio 1:2.6). After milling, the powders were freeze dried for 24 h at 50 C and 0.01 mbar (Labconco FreeZone 2.5) and sieved for 30 min to be separated from beads. Ce0.8Sm0.2O2-d powder (SDC20 e HP Fuel Cell Materials) was used for electrolyte preparation; it was cold pressed at 60 MPa, sintered in one step at 1450 C for 5 h in air, resulting in electrolyte disks with a diameter of 25 mm and a thickness of 1.2 mm. An ink for slurry coating of the electrode on the electrolyte disks was prepared by mixing the above powders in a mortar with alpha-therpineol (tech grade 90%, Sigma Aldrich; 50 drops in 15 g of powder). Cells were prepared in the threeelectrode configuration and symmetry between working electrode (WE) and counter electrode (CE) was achieved by using a home-made mask system. The reference electrode (RE) was placed on the WE side of the cell at a distance of 3 times the electrolyte thickness, as suggested in the literature [37,38] to avoid polarization problems during impedance measurements. All cathodes were sintered at 1100 C for 2 h, obtaining a geometric area of 0.28 cm2 and a thickness between 50 and 70 mm after sintering. In order to identify the crystalline phases and check the cathodes stability during operation, electrodes were analysed by X-ray diffraction (XRD, Cubix, Panalytical), using Cu-Ka incident radiation. Patterns were collected over the angular range 10 < 2q < 70 , in a constant scan mode with steps of 0.02 and a counting time of 7 s. XRD analysis was carried out on cathodes deposited on SDC electrolyte (half cells). Polished cross sections of same samples were also analysed by Scanning Electron Microscopy (SEM, 1450-vp, LEO) and by an energy-dispersive electron microprobe (EDX, INCA 300, Oxford Instruments). Electrochemical impedance spectroscopy (EIS) measurements were carried out in a three-electrode configuration to investigate the cathode catalytic activity; to this purpose, the cells were placed in a lab-manufactured test station [39,40]. A Pt mesh was placed on the surface of the WE and CE as current collector and Pt wires provided the connection to a potentiostat coupled to a frequency response analyser (Autolab PSSTAT302N). Impedance data were collected in the temperature range of 500e650 C at open circuit voltage (OCV) and at applied cathodic overpotentials (up to 0.3 V), with a perturbation of 10 mV in the 0.1e100 kHz frequency range. Ageing tests were performed at 650 C by applying a cathodic direct current (dc) of 200 mA cm2 and checking the system evolution by impedance measurements at OCV, after interruption of the dc load at different times. The cathode behavior was investigated through the analysis of the impedance spectra following two different approaches: equivalent circuits (ZView®, Scribner Associates Inc.) and distribution of relaxation time (DRT). DRT analysis is a deconvolution techniques, which allows one to point out the larger contribution to the frequency
6214
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
dispersion [41,42]. The distribution can be obtained by solving an equation of this type: Z∞ Zðui Þ ¼ R∞ þ Rp ∞
GðtÞ dlnt 1 þ jui
(1)
where Zðui Þ are the experimental data, R∞ is the resistance at high frequency, RP the polarization resistance and G(t) is a normalised function: Z∞ GðtÞdlnt ¼ 1
(2)
∞
Nevertheless this is an ill-posed inverse problem, and many different G(t) solutions are possible or the solution might be unstable, so that different approaches for the regularisation of the solution are available in literature [43e46]. In this paper the DRT technique was applied following Ciucci group's approach and tool [47].
Results and discussion Structural results Fig. 1a shows the XRD profile of BL30, BL50 and BL70 cathodes both as-sintered (AS) and after testing (AT). The profile of a 5050 v/v% mixture of starting powders is also reported for comparison. The structures of starting powders are confirmed as cubic (space group: Pm3m) for BSCF and trigonal (space group: R3c : H) for LSCF. Fig. 1b focuses on one angular range where no peaks from the substrate are visible. In general, both cubic and trigonal structures can be identified in the pattern of all samples, indicated in Fig. 1b as vertical lines. In AS samples, the peaks of the trigonal structure are slightly shifted (<0.08 deg) with respect to the LSCF, while for the cubic phase a significant shift with respect to those of BSCF is observed. The 2q shift is larger in BL30 (y1.2 deg) than in BL50 (y1.15 deg) and in BL70 (y0.8 deg). After testing, BL30 and BL50 are substantially unchanged. On the contrary, in BL70 the peaks of cubic phase are shifted back toward lower 2q angles, while the trigonal phase almost disappeared (see Fig. 1(b3)). These findings can be qualitatively explained by cationic interdiffusion between the two perovskite structures, taking place during sintering at 1100 C. As a result, the stoichiometry of each phase might deviate from the nominal and cannot be determined from standard XRD. Taking into account the incorporation of some smaller La3þ ions (r ¼ 136 pm, CN ¼ 12) replacing larger Sr2þ ions (r ¼ 144 pm, CN ¼ 12) at A-sites of BSCF, a unit cell volume decrease resulting in the positive shift of diffraction peaks position of the cubic structure might be expected. Similarly, the incorporation of Sr2þ replacing La3þ at A-sites of LSCF would cause lattice expansion, in this case balanced by shrinkage due to the formation of oxygen vacancies as charge compensation defect. The overall result is a small shift for the peaks of trigonal structure. By increasing the starting volume fraction of BSCF, the cubic structure prevails with the same interdiffusion and
Fig. 1 e (a) XRD profiles of samples, listed from base to top of diagram: BL70 (AS, dark cyan), BL70 (AT, magenta), BL50 (AS, cyan), BL50 (AT, blue), BL30 (AS, red) BL30 (AT, green), Mixture of starting powders (black). AS ¼ as sintered, AT ¼ after electrochemical test. (b) Magnification. (-) ¼ BSCF, (:) ¼ LSCF, (*) ¼ SDC, (d) ¼ cubic and (d. d) ¼ trigonal. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
incorporation phenomena taking place during sintering. On ageing, the observed backward shift of cubic phase peaks might be explained by progressive incorporation of La, i.e. formation of BLSCF, compensated by elimination of oxygen vacancies, causing lattice expansion. SEM analyses showed that BL30 and BL50 electrode layers were adherent to the electrolyte substrate and with regular shape and thickness, in the range 50÷60 mm. The two phases appeared uniformly mixed although some areas with finer grain size, about 0.4 mm diameter, and others organised in larger clusters, about 2e4 mm diameter, can be observed. Given the particle size and shape of starting
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
powders, i.e. 200 nm spherical for LSCF and about 3 mm faceted for BSCF [34], the two different morphologies observed can be respectively ascribed to LSCF-like or BSCF-like structure. Coherently, the relative amount of larger clusters increased with starting amount of BSCF. These features were not affected by electrochemical testing and ageing and are clearly visible in Fig. 2a and b. It is worth noting that the mean atomic number of La0.6Sr0.4Co0.2Fe0.8O3 and Ba0.5Sr0.5Co0.8Fe0.2O3, calculated by the Bu¨chner formula [48], are very close and therefore will appear with the same grey tone in back scattered electron (BSE) images. Sample BL70 showed good adhesion to the substrate (Fig. 2c), with irregular thickness, ranging between 50 and 70 mm. The microstructure is characterized by the presence of large clusters (>30 mm) and pores with diameter in the 2e10 mm range. After testing the electrode layer was adherent to the electrolyte. Thickness was reduced, between 30 and 40 mm, due to further growth of clusters, with also formation of large voids (Fig. 2d). No secondary phases were detected and the composition obtained by EDX was quite homogeneous within each electrode and corresponding to the average nominal one. However, the formation of small amounts of secondary phases, below the detection limit of the technique and instrumentation used, cannot be ruled out in all samples.
Electrochemical results Electrochemical impedance spectroscopy investigation (EIS) was carried out in three-electrode configuration considering a temperature range between 500 and 650 C (DT ¼ 25 C). Fig. 3a shows the typical impedance spectra for each composition at 650 C in air, compared to reference materials. In these
6215
operating conditions, the best performances were obtained by BL70 and BL50, with a RP equal to 0.021 and 0.027 U cm2, respectively. Nevertheless, also BL30 mixture provided a significant improvement of electrode activity with a RP value equal to 0.045 U cm2; indeed, for starting pure commercial LSCF and BSCF electrodes, measured RP were, 0.46 U cm2 and 0.21 U cm2, respectively. All the RP values obtained for new electrodes are competitive if compared with recent values reported in literature for optimized BSCF-based electrodes [49], which still remains the most active perovskite-based SOFCs cathode [2]. Activation energy (Eact) values calculated from Arrhenius plots are reported in Fig. 3b. In spite of its higher resistivity, BL30 shows the lowest Eact (0.96eV), indicating that this composition becomes more competitive at low temperatures. The sample BL70 has an activation energy value of 1.17 eV close to the pure BSCF (1.18 eV) and lower than pure LSCF (1.52 eV), in agreement with those reported in literature (LSCF 1.45 eV [50], BSCF 1.20 eV [20,51]). The Eact values confirm that all the three new compositions are prone to oxygen reduction reaction, with an enhanced electrocatalytic activity. As reported in Fig. 4a, showing the Nyquist plots of sample BL70 as an example, a remarkable difference in the impedance shape was observed according to the temperature. This strong effect was detected for all the compositions and it indicates that not only polarization resistance, but also the whole kinetic mechanism is very sensitive to this operating parameter, as confirmed by the analysis of distribution of relaxation time (DRT), performed on EIS data at OCV, and illustrated in Fig. 4b. This not trivial behavior prevented the use of a unique and reliable equivalent circuit able to describe all the experimental spectra in the whole temperature range. On the other hand,
Fig. 2 e BSE images on cross section of samples: (a) BL30 as sintered, (b) BL50 after electrochemical testing, (c) BL70 as sintered and (d) BL70 after electrochemical testing.
6216
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Fig. 3 e (a) Impedance spectra at 650 C and OCV for BL30, BL50, BL70. The inset shows impedance spectra for pure BSCF and LSCF. (b) Arrhenius plot for the five cathodes and calculated activation energy: BL30 (0.95 eV), BL50 (1.06 eV), BL70 (1.17 eV), BSCF (1.18 eV) and LSCF (1.52 eV).
Fig. 4 e (a) Impedance spectra and (b) corresponding DRT patterns obtained for sample BL70 at OCV, in the range 500e650 C. The inset in (b) shows a magnification of the DRT curves at 600 and 650 C. distribution of relaxation time (DRT) allowed recognition of the change in predominant phenomena for the two different regimes (500 and 650 C). DRT spectra at 500 C showed a main peak at low frequencies (about 10 Hz, Fig. 4b), followed by a tail at mediumhigh frequencies. This is a typical Gerischer behavior, as reported in different studies [52,53], which implies considering the system as co-controlled by coupled oxygen surface exchange and bulk diffusion, according to the following expression [54]: ZG ¼ Rchem
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ jutchem
(3)
Rchem being the chemical resistance and tchem the characteristic time, strictly related to oxygen surface exchange and transport properties. Positive feedback for the Gerischer hypothesis came from the analysis of impedance data through an R1G equivalent circuit fitting, as illustrated in Fig. 5. Increasing the temperature (typically at 600e650 C) the main peak shifted to a medium-high frequency range (around 140 Hz, Fig. 4b) and its intensity significantly decreased,
Fig. 5 e BL30 and BL70 cathode impedance spectra at 500 C and OCV overlaid on the fitting spectra. R1 ¼ (subtracted) electrolyte resistance; GE1 ¼ Gerischer element (named G in the text).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
indicating a thermal activation of the process and a possible kinetic mechanism modification. Furthermore, at these higher temperatures a second well-defined peak appeared at lower frequency (around 1 Hz, Fig. 4b). Based on the peak frequency and data reported in literature, it is reasonable to consider this process related to surface phenomena, such as oxygen adsorption, or to oxygen diffusion at gas phase [1,55,56]. Going more into detail, at 650 C the total resistance decreases compared to 600 C (Fig. 4a), but the area under low frequency peak (strictly linked to resistance phenomena) rises (inset in Fig. 4b), and its maximum value (PLF) moves to lower frequencies, indicating a process hindered by the temperature increasing. The same temperature effect at low frequency was detected by Amin and Karan [57] for a La0.5Ba0.5CoO3-based cathode; they suggested that this behavior it could be due to an unfavorable oxygen adsorption on the electrode surface, or a partial surface coverage at higher temperature. Nevertheless, the identification of the process in this frequency range is not trivial due to the complexity of the system and the multiplicity of phenomena involved, as also reported by Bidrawn et al. [58], it is extremely difficult to distinguish whether diffusion or oxygen adsorption is the rate-limiting step. Considering the kinetic regime change, further investigation was carried out under cathodic polarization condition in the studied range of temperature. Fig. 6 reports the impedance spectra at different cathodic overpotentials (h) at 500 and 650 C,
6217
with the matched distribution of relaxation times (insets) both for BL30 (Fig. 6a and b) and for BL70 sample (Fig. 6c and d). According to Adler model [54], by manipulating relationship (3), it is possible to obtain the imaginary part of impedance from: vq ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 u ðu$t chem Þ þ 1 1 Rchem t Z''G ¼ pffiffiffi 2 ðu$tchem Þ2 þ 1
(4)
and then extrapolating tchem and Rchem fitting experimental data. Their ratio allows to calculate a chemical capacitance: Cchem ¼ tchem =Rchem
(5)
useful to evaluate the electrode bulk ability to store charge in terms of oxygen vacancies. For both samples, at 500 C, Rchem is a good approximation of global RP for all the applied h, confirming the dominant Gerischer behavior introduced above. More into detail, at 500 C, BL30 exhibits the same behavior observed by the authors for LSCF reference electrode, with a RP reduction as h increases. A possible explanation of this behavior, already proposed for LSCF [26], is the increase on ð$$Þ oxygen vacancy concentration [VO ] under cathodic h and their fundamental contribution to oxygen solid-state diffuð$$Þ sion through the electrode bulk. This [VO ] variation is highlighted also by Cchem response (Fig. 7b) that grows under low h.
Fig. 6 e Impedance spectra and matched distribution of relaxation times at 500 and 600 C for sample BL30 (a and b) and for BL70 (c and d) sample. The mean peak at 500 C appears at 11e12 Hz ((a) and (c)); the mean peak at 650 C appears at about 140 Hz ((b) and (d)).
6218
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Fig. 7 e (a) Rchem, and (b) Cchem as functions of applied overpotential at 500 C, derived from Gerischer fitting, for BL30 and BL70.
Although BL30 shows a smooth Cchem decrease at higher h, its final value globally increases if compared with that calculated in OCV conditions. For BL70 electrode h has an opposite effect (Rp increases with h), similar to BSCF e behavior investigated by some of the authors in previous paper [26]. This work and other papers [34,59,60] highlighted that an applied overpotential can cause ð$$Þ oxygen vacancy cluster formation as [VO ] increases, with consequent losses for oxygen reduction activity of the electrode, explaining why the Cchem, starting increase is not able to lead an improvement in electrode activity. When the system is brought at 650 C the more resistive process was located at medium-high frequency (Fig. 6b and d) pointing out a kinetic change compared to 500 C behavior. Based on DRT analysis and on a previous study [34], the circuit R1 (RHFQHF) (RLFQLF) was used to fit the impedance results. In this equivalent circuit R1 is the electrolyte resistance, RHF and QHF are the resistance and constant phase element (CPE) at high frequency (~140 Hz) and RLF and QLF describe the system at low frequency. According to Ref. [61] a capacitance was
calculated from CPE parameters. All these parameters are reported in Fig. 8a and b, which show that the limiting step at 650 C is the high frequency process (Fig. 6b and d). Also at 650 C, according to their high values, the low frequency capacitance is chemical for both samples, confirming the bulk ability to store charge as oxygen vacancies. This ability is enhanced in composites and increases with BSCF content, as indicated by Cchem values at 650 C and OCV: in fact, Cchem is equal to 0.068 in LSCF, 0.806 in BSCF, raising in BL30, BL50 and BL70 up to 1.14, 5.2 and 12.9 respectively (values expressed in F.cm2). The above-described regime change in the range 500e650 C represents the key to explain the improvement ensured by composite over single-phase cathodes. In fact, remarkable performances of LSCF and BSCF composites are a due to drastic reduction of the polarization resistance associated with oxygen surface exchange and bulk diffusion. This phenomenon can be considered a combined effect of the high oxygen vacancies concentration in BSCF and of the high ionic conductivity in LSCF. The main limitation to conduction at 650 C is still represented by charge transfer at electrodeelectrolyte interface, which could be further improved by electrode fabrication process optimization. After the electrochemical investigation, an ageing test at 650 C under a current load of 200 mA/cm2 was performed to evaluate electrode degradation over the time (Fig. 9). Degradation of electrode performance in the first operating hours is reported as key factor for cathode materials life. Analysis of performance losses in the first 100e300 working hours has been taken as good indication to evaluate electrode stability for longer operating time [62e64]. Consequently, an ageing time of 200 h was chosen to monitor the evolution of the system performing periodic EIS measurements at regular time. All the electrodes showed an initial remarkable degradation, but Rp trend are different for the three compositions. BL30 and BL50 showed a quite linear increase of Rp, while for BL70 after the first 50 h the degradation rate slowed down, remaining the most attractive cathode. The Rp value stabilized close to 0.048 U cm2, that is highly competitive even if compared with resistance data of pure and fresh LSCF and BSCF cathode reported in literature [31,52,55,63,65e71] and
Fig. 8 e Values of resistances (a) and capacitances (b) extracted from the fitting analysis of impedance results obtained at 650 C for BL30 and BL70 samples.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Fig. 9 e Degradation rate for BL30, BL50, BL70 electrodes and BL70 chemical capacitance evolution under an applied current load of 200 mA/cm2 at 650 C.
also with highly-performing BLSCF [63]. Confirmation about stabilization of BL70 is provided by the decrease of degradation rate; indeed, three different intervals were identified between 0 and 24 h, 24e72 h and above 72 h with a slope of 6.67 104, 1.66 104 and 3.58 105 U cm2 h1 respectively. The starting order of magnitude for degradation rate (~104 U cm2 h1) is in accordance with those extrapolated for BSCF and LSCF (103 ÷104 U cm2 h1) by the authors and other papers [26,34,62], while the drop to ~105 U cm2 h1 (R2 ¼ 0.992) represents a stability improvement for electrode systems based on perovskite materials. XRD and SEM investigations did not show structural and morphological modifications during ageing of BL30 and BL50 samples. It is known [72] that, at SOFC operative temperature, segregation of Sr and Co is likely to take place in LSCF. Therefore, the formation of tiny layers of Sr oxides with limited conductivity can be expected at grain boundaries on aging, but cannot be detected due to the resolution of standard SEM and XRD used in this work as reported in the structural analysis section. The increase of Rp (Fig. 9) with aging time in BL30 and BL50 can be ascribed to this phenomenon, which is more pronounced in BL30, i.e. the mixture with highest LSCF content. On the other hand, the observed backward shift of cubic phase XRD peaks during BL70 ageing, explained as progressive incorporation of La compensated by a decrease of oxygen vacancies, find confirmation on the measure of the amount of oxygen vacancies through CChem, which shows a decrease from 10 to 1 F cm2 in the first 50 h followed by a stable value in the remaining hours. Such changes in the bulk of the electrode match well with the BL70 performance losses displayed in the first 50 h of ageing of Fig. 9 and the morphology evolution observed by SEM analysis. As a matter of fact, BL30 and BL50 CChem remain unchanged during ageing, indicating a substantial constant concentration of oxygen vacancies.
Conclusions The effect of composition in performance and stability of La0.6Sr0.4Co0.2Fe0.8O3-d (LSCF) - Ba0.5Sr0.5Co0.8Fe0.2O3-d (BSCF)
6219
based cathodes was investigated. The mixtures were prepared through ball milling of commercial LSCF and BSCF powders. The cathodes were deposited on Ce0.8Sm0.2O2-d-electrolytes and investigated in SOFC typical conditions of operation (different temperatures and cathodic overpotentials). Microstructural investigation revealed cations interdiffusion between the cubic BSCF and trigonal LSCF structures during sintering, without formation of secondary phases. For BL70, the cubic structure was predominant and showed lattice parameter increase on ageing, compatible with oxygen uptake from atmosphere. Analyses of the impedance results by distribution of relaxation time (DRT) and equivalent circuits showed two kinetic regimes governing the electrocatalytic activity. At low temperature (500 C), the more resistive step was identified at low frequencies and global impedance was fitted through a Gerischer element, pointing out a behavior co-controlled by surface oxygen exchange and bulk diffusion; this model was no longer suitable for the high temperature regime (650 C). The resistance trends for BL30 and BL70 as functions of overpotential appeared different: an increase of h led to a Rchem decrease for BL30 (LSCF-like behavior) and to an increase for BL70 sample (BSCF-like behavior). Like BSCF, also BL70 electrochemical activity decreased under applied overpotential, because of a suppression of vacancy concentration, and this behavior was not mitigated by the presence of a significant (30%) starting amount of LSCF. All cathodes showed very good electrochemical performance. A preliminary ageing test (200 h) was carried out at 650 C; despite the observed degradation for all three investigated cathodes, BL70 Rp showed an almost asymptotic trend, displaying, for the ageing studied time, a final stabilized and low value close to 0.048 U cm2. These first results open the way for a further optimization step to modulate physical and chemical electrode properties, with a more detailed experimental approach based on application of strategies already available to improve electrode stability and activity. Nevertheless, investigations at longer loading time are necessary and planned for the near future, especially for BL70 composition.
references
[1] Adler SB. Factors governing oxygen reduction in solid oxide fuel cell cathodes. Chem Rev 2004;104:4791e844. https:// doi.org/10.1021/cr020724o. [2] Areum J, Junyoung K, Jeeyoung S, Guntae K. Perovskite as a cathode material: a review of its role in solid-oxide fuel cell technology. ChemElectroChem 2016;3:511e30. https:// doi.org/10.1002/celc.201500382. [3] Carpanese MP, Barbucci A, Canu G, Viviani M. BaCe0.85Y0.15O2.925 dense layer by wet powder spraying as electrolyte for SOFC/SOEC applications. Solid State Ionics 2015;269:80e5. https://doi.org/10.1016/j.ssi.2014.11.014. [4] Skinner SJ. Recent advances in Perovskite-type materials for solid oxide fuel cell cathodes. Int J Inorg Mater 2001;3:113e21. https://doi.org/10.1016/S1466-6049(01)00004-6. [5] Viviani M, Canu G, Carpanese MP, Barbucci A, Sanson A, Mercadelli E, et al. Dual cells with mixed protonic-anionic conductivity for reversible SOFC/SOEC operation. Energy
6220
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
Procedia 2012;28:182e9. https://doi.org/10.1016/j.egypro. 2012.08.052. Vladikova D, Stoynov Z, Chesnaud A, Thorel A, Viviani M, Barbucci A, et al. Application of yttrium doped barium cerate for improvement of the dual membrane SOFC design. Int J Hydrogen Energy 2014;39:21561e8. https://doi.org/10.1016/j. ijhydene.2014.09.014. da Silva FS, de Souza TM. Novel materials for solid oxide fuel cell technologies: a literature review. Int J Hydrogen Energy 2017;42:26020e36. https://doi.org/10.1016/j.ijhydene. 2017.08.105. Dai Y, Lou Z, Wang Z, Qiao J, Sun W, Sun K. Electrochemical and chemical stability performance improvement of Ba0.5Sr0.5Fe0.91Al0.09O3d cathode for IT-SOFC through the introduction of a GDC interlayer. Int J Hydrogen Energy 2015;40:5939e46. https://doi.org/10.1016/j.ijhydene.2015.02.131. Huang X, Feng J, Abdellatif HRS, Zou J, Zhang G, Ni C. Electrochemical evaluation of double perovskite PrBaCo2xMnxO5þd (x ¼ 0, 0.5, 1) as promising cathodes for IT-SOFCs. Int J Hydrogen Energy 2018;43:8962e71. https://doi.org/10. 1016/j.ijhydene.2018.03.163. Nicollet C, Flura A, Vibhu V, Rougier A, Bassat J-M, Grenier JC. An innovative efficient oxygen electrode for SOFC: Pr6O11 infiltrated into Gd-doped ceria backbone. Int J Hydrogen Energy 2016;41:15538e44. https://doi.org/10.1016/j.ijhydene. 2016.04.024. pez C, Hauch A, Ruiz de Ortiz-Vitoriano N, Bernuy-Lo Larramendi I, Rojo T. Electrochemical characterization of La0.6Ca0.4Fe0.8Ni0.2O3 cathode on Ce0.8Gd0.2O1.9 electrolyte for IT-SOFC. Int J Hydrogen Energy 2014;39:6675e9. https://doi. org/10.1016/j.ijhydene.2014.01.032. Mineshige A, Izutsu J, Nakamura M, Nigaki K, Abe J, Kobune M, et al. Introduction of A-site deficiency into La0.6Sr0.4Co0.2Fe0.8O3ed and its effect on structure and conductivity. Solid State Ionics 2005;176:1145e9. https://doi. org/10.1016/j.ssi.2004.11.021. Kiebach R, Zhang W-W, Zhang W, Chen M, Norrman K, Wang H-J, et al. Stability of La0.6Sr0.4Co0.2Fe0.8O3/Ce0.9Gd0.1O2 cathodes during sintering and solid oxide fuel cell operation. J Power Sources 2015;283:151e61. https://doi.org/10.1016/j. jpowsour.2015.02.064. Wang F, Nishi M, Brito ME, Kishimoto H, Yamaji K, Yokokawa H, et al. Sr and Zr diffusion in LSCF/10GDC/8YSZ triplets for solid oxide fuel cells (SOFCs). J Power Sources 2014;258:281e9. https://doi.org/10.1016/j.jpowsour.2014.02. 046. Wang H, Yakal-Kremski KJ, Yeh T, Rupp GM, Limbeck A, Fleig J, et al. Mechanisms of performance degradation of (La,Sr)(Co,Fe)O3-d solid oxide fuel cell cathodes. J Electrochem Soc 2016;163:F581e5. https://doi.org/10.1149/2.0031607jes. Khan MZ, Mehran MT, Song R-H, Lee J-W, Lee S-B, Lim T-H, et al. Effect of GDC interlayer thickness on durability of solid oxide fuel cell cathode. Ceram Int 2016;42:6978e84. https:// doi.org/10.1016/j.ceramint.2016.01.085. Wang F, Brito ME, Yamaji K, Cho D-H, Nishi M, Kishimoto H, et al. Effect of polarization on Sr and Zr diffusion behavior in LSCF/GDC/YSZ system. Solid State Ionics 2014;262:454e9. https://doi.org/10.1016/j.ssi.2014.04.002. Khan MZ, Mehran MT, Song R-H, Lee S-B, Lim T-H. Effects of applied current density and thermal cycling on the degradation of a solid oxide fuel cell cathode. Int J Hydrogen Energy 2018;43:12346e57. https://doi.org/10.1016/j.ijhydene. 2018.04.175. Araki W, Arai Y, Malzbender J. Transitions of Ba0.5Sr0.5Co0.8Fe0.2O3-d and La0.58Sr0.4Co0.2Fe0.8O3-d. Mater Lett 2014;132:295e7. https://doi.org/10.1016/j.matlet.2014.06.093. Shao Z, Haile SM. A high-performance cathode for the next generation of solid-oxide fuel cells. Nature 2004;431:170.
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
https://doi.org/10.1038/nature02863. https://www.nature. com/articles/nature02863#supplementary-information. Mueller DN, De Souza RA, Weirich TE, Roehrens D, Mayer J, Martin M. A kinetic study of the decomposition of the cubic perovskite-type oxide BaxSr1-xCo0.8Fe0.2O3-d (BSCF) (x ¼ 0.1 and 0.5). Phys Chem Chem Phys 2010;12:10320e8. https:// doi.org/10.1039/c0cp00004c. ng Z, Martynczuk J, Efimov K, Harvey AS, Infortuna A, Ya Kocher P, et al. Oxygen-vacancy-related structural phase transition of Ba0.8Sr0.2Co0.8Fe0.2O3-d. Chem Mater 2011;23:3169e75. https://doi.org/10.1021/cm200373r. Bermudez JM, Garcia-Fayos J, Reina TR, Reed G, Persoon ES, € rtz D, et al. Thermochemical stability of LaxSr1-xCoyFe1Go yO3-d and NiFe2O4-Ce0.8Tb0.2O2-d under real conditions for its application in oxygen transport membranes for oxyfuel combustion. J Membr Sci 2018;562:26e37. https://doi.org/10. 1016/j.memsci.2018.05.012. Arnold M, Wang H, Feldhoff A. Influence of CO2 on the oxygen permeation performance and the microstructure of perovskite-type (Ba0.5Sr0.5)(Co0.8Fe0.2)O3d membranes. J Membr Sci 2007;293:44e52. https://doi.org/10.1016/j.memsci. 2007.01.032. Zhang Y, Yang G, Chen G, Ran R, Zhou W, Shao Z. Evaluation of the CO2 poisoning effect on a highly active cathode SrSc0.175Nb0.025Co0.8O3-d in the oxygen reduction reaction. ACS Appl Mater Interfaces 2016;8:3003e11. https://doi.org/ 10.1021/acsami.5b09780. Giuliano A, Carpanese MP, Clematis D, Boaro M, Pappacena A, Deganello F, et al. Infiltration, overpotential and ageing effects on cathodes for solid oxide fuel cells: La0.6Sr0.4Co0.2Fe0.8O3-d versus Ba0.5Sr0.5Co0.8Fe0.2O3-d. J Electrochem Soc 2017;164:F3114e22. https://doi.org/10.1149/ 2.0161710jes. Ai N, Chen K, Jiang SP. A La0.8Sr0.2MnO3/ La0.6Sr0.4Co0.2Fe0.8O3d coreeshell structured cathode by a rapid sintering process for solid oxide fuel cells. Int J Hydrogen Energy 2017;42:7246e51. https://doi.org/10.1016/j. ijhydene.2016.10.036. Chen J, Wan D, Sun X, Li B, Lu M. Electrochemical impedance spectroscopic characterization of impregnated La0.6Sr0.4Co0.2Fe0.8O3d cathode for intermediate-temperature SOFCs. Int J Hydrogen Energy 2018;43:9770e6. https://doi.org/ 10.1016/j.ijhydene.2018.03.223. Naumovich EN, Zakharchuk K, Obre˛bowski S, Yaremchenko A. (La,Sr)(Fe,Co)O3-based cathode contact materials for intermediate-temperature solid oxide fuel cells. Int J Hydrogen Energy 2017;42:29443e53. https://doi.org/10. 1016/j.ijhydene.2017.09.122. Zhou F, Liu Y, Zhao X, Tang W, Yang S, Zhong S, et al. Effects of cerium doping on the performance of LSCF cathodes for intermediate temperature solid oxide fuel cells. Int J Hydrogen Energy 2018;43:18946e54. https://doi.org/10.1016/j. ijhydene.2018.08.041. Zhang X, Zhang W, Zhang L, Meng J, Meng F, Liu X, et al. Enhanced electrochemical property of La0.6Sr0.4Co0.8Fe0.2O3 as cathode for solid oxide fuel cell by efficient in situ polarization-exsolution treatment. Electrochim Acta 2017;258:1096e105. https://doi.org/10.1016/j.electacta.2017. 11.163. Perry Murray E, Sever MJ, Barnett SA. Electrochemical performance of (La,Sr)(Co,Fe)O3e(Ce,Gd)O3 composite cathodes. Solid State Ionics 2002;148:27e34. https://doi.org/ 10.1016/S0167-2738(02)00102-9. Liu Y, Bi J, Chi B, Pu J, Jian L. Effects of impregnating palladium on catalytic performance of LSCF-GDC composite cathodes for intermediate temperature solid oxide fuel cells. Int J Hydrogen Energy 2016;41:6486e92. https://doi.org/10. 1016/j.ijhydene.2016.02.142.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
[34] Giuliano A, Carpanese MP, Panizza M, Cerisola G, Clematis D, Barbucci A. Characterisation of La0.6Sr0.4Co0.2Fe0.8O3-d e Ba0.5Sr0.5Co0.8Fe0.2O3-d composite as cathode for solid oxide fuel cells. Electrochim Acta 2017;240:258e66. https://doi.org/ 10.1016/j.electacta.2017.04.079. ska E, Tatko M. [35] Mosiałek M, Ke˛dra A, Krzan M, Bielan Ba0.5Sr0.5Co0.8Fe0.2O3edeLa0.6Sr0.4Co0.8Fe0.2O3ed composite cathode for solid oxide fuel cell. Arch Metall Mater 2016;61. https://doi.org/10.1515/amm-2016-0243. [36] Kim J-S, Yeon D-H, Jung DW, Kwak C. A highly active and long-term stable La-doped BaxSr1xCo1yFeyO3d cathode for solid-oxide fuel cells. J Power Sources 2014;249:66e71. https://doi.org/10.1016/j.jpowsour.2013.10.053. [37] Cimenti M, Co AC, Birss VI, Hill JM. Distortions in electrochemical impedance spectroscopy measurements using 3-electrode methods in SOFC. I e Effect of cell geometry. Fuel Cells 2007;7:364e76. https://doi.org/10.1002/ fuce.200700019. [38] Winkler J, Hendriksen PV, Bonanos N, Mogensen M. Geometric requirements of solid electrolyte cells with a reference electrode. J Electrochem Soc 1998;145:1184e92. https://doi.org/10.1149/1.1838436. [39] Carpanese MP, Panizza M, Viviani M, Mercadelli E, Sanson A, Barbucci A. Study of reversible SOFC/SOEC based on a mixed anionic-protonic conductor. J Appl Electrochem 2015;45:657e65. https://doi.org/10.1007/s10800-015-0838-8. [40] Presto S, Barbucci A, Carpanese MP, Viviani M, Marazza R. Electrochemical performance of Ni-based anodes for solid oxide fuel cells. J Appl Electrochem 2009;39:2257e64. https:// doi.org/10.1007/s10800-009-9857-7. [41] Boukamp BA, Rolle A. Analysis and application of distribution of relaxation times in solid state ionics. Solid State Ionics 2017;302:12e8. https://doi.org/10.1016/j.ssi.2016. 10.009. e E. [42] Schichlein H, Mu¨ller AC, Voigts M, Kru¨gel A, Ivers-Tiffe Deconvolution of electrochemical impedance spectra for the identification of electrode reaction mechanisms in solid oxide fuel cells. J Appl Electrochem 2002;32:875e82. https:// doi.org/10.1023/a:1020599525160. € rlin T. Deconvolution and maximum entropy in [43] Ho impedance spectroscopy of noninductive systems. Solid State Ionics 1998;107:241e53. https://doi.org/10.1016/S01672738(98)00008-3. [44] Renaut RA, Baker R, Horst M, Johnson C, Nasir D. Stability and error analysis of the polarization estimation inverse problem for microbial fuel cells. Inverse Probl 2013;29:045006. € nleber M, Weber A, Ivers-Tiffe e E. [45] Schmidt JP, Berg P, Scho The distribution of relaxation times as basis for generalized time-domain models for Li-ion batteries. J Power Sources 2013;221:70e7. https://doi.org/10.1016/j.jpowsour.2012.07. 100. [46] Tuncer E, Macdonald JR. Comparison of methods for estimating continuous distributions of relaxation times. J Appl Phys 2006;99:074106. https://doi.org/10.1063/1.2188053. [47] Wan TH, Saccoccio M, Chen C, Ciucci F. Influence of the discretization methods on the distribution of relaxation times deconvolution: implementing radial basis functions with DRTtools. Electrochim Acta 2015;184:483e99. https:// doi.org/10.1016/j.electacta.2015.09.097. [48] Howell PGT, Davy KMW, Boyde A. Mean atomic number and backscattered electron coefficient calculations for some materials with low mean atomic number. Scanning 1998;20:35e40. https://doi.org/10.1002/sca.1998.4950200105. € rmer H, Meffert M, Sza sz J, Wankmu¨ller F, [49] Almar L, Sto Gerthsen D, et al. Improved phase stability and CO2 poisoning robustness of Y-doped Ba0.5Sr0.5Co0.8Fe0.2O3d SOFC cathodes at intermediate temperatures. ACS Appl
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
6221
Energy Mater 2018;1:1316e27. https://doi.org/10.1021/ acsaem.8b00028. Serra JM, Garcia-Fayos J, Baumann S, Schulze-Ku¨ppers F, Meulenberg WA. Oxygen permeation through tape-cast asymmetric all-La0.6Sr0.4Co0.2Fe0.8O3d membranes. J Membr Sci 2013;447:297e305. https://doi.org/10.1016/j.memsci.2013. 07.030. Duan C, Hook D, Chen Y, Tong J, O'Hayre R. Zr and Y codoped perovskite as a stable, high performance cathode for solid oxide fuel cells operating below 500 C. Energy Environ Sci 2017;10:176e82. https://doi.org/10.1039/c6ee01915c. sz J, Weber A, Ivers-Tiffe e E. Oxygen transport Almar L, Sza kinetics of mixed ionic-electronic conductors by coupling focused ion beam tomography and electrochemical impedance spectroscopy. J Electrochem Soc 2017;164:F289e97. https://doi.org/10.1149/2.0851704jes. Zhang Y, Chen Y, Yan M, Chen F. Reconstruction of relaxation time distribution from linear electrochemical impedance spectroscopy. J Power Sources 2015;283:464e77. https://doi.org/10.1016/j.jpowsour.2015.02.107. Adler SB, Lane JA, Steele BCH. Electrode kinetics of porous mixed-conducting oxygen electrodes. J Electrochem Soc 1996;143:3554e64. https://doi.org/10.1149/1.1837252. s D, Soldati AL, Troiani H, Montenegro-Herna ndez A, Garce Caneiro A, Mogni LV. La/Ba-based cobaltites as IT-SOFC cathodes: a discussion about the effect of crystal structure and microstructure on the O2-reduction reaction. Electrochim Acta 2016;215:637e46. https://doi.org/10.1016/j. electacta.2016.08.132. Im H-N, Choi M-B, Jeon S-Y, Singh B, Song S-J. Oxygen reduction properties of La0.1Sr0.9Co0.8Fe0.2O3-d cathode for SOFC using electrochemical method. Meet Abstr 2014:729. MA2014-01. Amin R, Karan K. Characterization of La0.5Ba0.5CoO3 d as a SOFC cathode material. J Electrochem Soc 2010;157:B285e91. https://doi.org/10.1149/1.3267879. Bidrawn F, Ku¨ngas R, Vohs JM, Gorte RJ. Modeling impedance response of SOFC cathodes prepared by infiltration. J Electrochem Soc 2011;158:B514e25. https://doi.org/10.1149/ 1.3565174. Gangopadhyay S, Masunov AE, Inerbaev T, Mesit J, Guha RK, Sleiti AK, et al. Understanding oxygen vacancy migration and clustering in barium strontium cobalt iron oxide. Solid State Ionics 2010;181:1067e73. https://doi.org/10.1016/j.ssi.2010.05. 014. Tsipis EV, Naumovich EN, Patrakeev MV, Yaremchenko AA, Marozau IP, Kovalevsky AV, et al. Oxygen deficiency, vacancy clustering and ionic transport in (La,Sr)CoO3d. Solid State Ionics 2011;192:42e8. https://doi.org/10.1016/j.ssi.2010.03.029. Hirschorn B, Orazem ME, Tribollet B, Vivier V, Frateur I, Musiani M. Determination of effective capacitance and film thickness from constant-phase-element parameters. Electrochim Acta 2010;55:6218e27. https://doi.org/10.1016/j. electacta.2009.10.065. e E. LongEndler C, Leonide A, Weber A, Tietz F, Ivers-Tiffe Term study of MIEC cathodes for intermediate temperature solid oxide fuel cells. ECS trans 2009;25:2381e90. https:// doi.org/10.1149/1.3205791. Junyoung K, Sihyuk C, Areum J, Young JH, Jeeyoung S, Guntae K. Chemically stable perovskites as cathode materials for solid oxide fuel cells: La-doped Ba0.5Sr0.5Co0.8Fe0.2O3d. ChemSusChem 2014;7:1669e75. https://doi.org/10.1002/cssc.201301401. Wang WG, Mogensen M. High-performance lanthanumferrite-based cathode for SOFC. Solid State Ionics 2005;176:457e62. https://doi.org/10.1016/j.ssi.2004.09.007. Chen C-H, Chang C-L, Hwang B-H. Electrochemical and microstructure characteristics of Ba0.5Sr0.5Co0.8Fe0.2O3d
6222
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 4 ( 2 0 1 9 ) 6 2 1 2 e6 2 2 2
(BSCF) cathodes prepared by citrate precursor method for SOFCs. Mater Chem Phys 2009;115:478e82. https://doi.org/10. 1016/j.matchemphys.2009.01.018. [66] Chen Y, Qian B, Li S, Jiao Y, Tade MO, Shao Z. The influence of impurity ions on the permeation and oxygen reduction properties of Ba0.5Sr0.5Co0.8Fe0.2O3d perovskite. J Membr Sci 2014;449:86e96. https://doi.org/10.1016/j.memsci.2013.08.021. [67] Deganello F, Liotta LF, Marcı` G, Fabbri E, Traversa E. Strontium and iron-doped barium cobaltite prepared by solution combustion synthesis: exploring a mixed-fuel approach for tailored intermediate temperature solid oxide fuel cell cathode materials. Mater Renew Sustain Energy 2013;2:8. https://doi.org/10.1007/s40243-013-0008-z. MO, Shao Z. A top-down strategy [68] Su C, Xu X, Chen Y, Liu Y, Tade for the synthesis of mesoporous Ba0.5Sr0.5Co0.8Fe0.2O3d as a cathode precursor for buffer layer-free deposition on stabilized zirconia electrolyte with a superior electrochemical performance. J Power Sources 2015;274:1024e33. https://doi. org/10.1016/j.jpowsour.2014.10.177.
[69] Sumi H, Ohshiro T, Nakayama M, Suzuki T, Fujishiro Y. Prevention of reaction between (Ba,Sr)(Co,Fe)O3 cathodes and yttria-stabilized zirconica electrolytes for intermediate-temperature solid oxide fuel cells. Electrochim Acta 2015;184:403e9. https://doi.org/10.1016/j.electacta.2015. 10.092. [70] Wang K, Ran R, Zhou W, Gu H, Shao Z, Ahn J. Properties and performance of Ba0.5Sr0.5Co0.8Fe0.2O3dþSm0.2Ce0.8O1.9 composite cathode. J Power Sources 2008;179:60e8. https:// doi.org/10.1016/j.jpowsour.2007.12.051. [71] Zhu X, Xia H, Li Y, Lu¨ Z. A (La,Sr)MnO3 nano-film embedded into (Ba,Sr)(Co,Fe)O3 porous cathode for stability enhancement. Mater Lett 2015;161:549e53. https://doi.org/ 10.1016/j.matlet.2015.09.040. [72] Hardy JS, Coyle CA, Bonnett JF, Templeton JW, Canfield NL, Edwards DJ, et al. Evaluation of cation migration in lanthanum strontium cobalt ferrite solid oxide fuel cell cathodes via in-operando X-ray diffraction. J Mater Chem 2018;6:1787e801. https://doi.org/10.1039/c7ta06856e.