Physicochemical properties of ceramic tape involving Ca0.05 Ba0.95 Ce0.9Y0.1O3 as an electrolyte designed for electrolyte-supported solid oxide fuel cells (IT-SOFCs)
- 181 Downloads
A proton-conducting membrane involving modified barium cerate BaCe0.9Y0.1O3 was obtained in the form of gas-tight ceramic tape. Monophase Ca0.05Ba0.95Ce0.9Y0.1O3 (5CBCY) powder and an organic medium consisting of polyvinyl butyral used as a binder, a plasticiser based on carboxylic acid esters, and a mixture of ethanol and methyl ethyl ketone were used to prepare slurry for the tape-casting process. Gas-tight ceramic tapes involving 5CBCY and sintered samples were tested as electrolytes in hydrogen–oxygen button solid oxide fuel cells within the temperature range 500–750 °C. Variations in OCV and ohmic resistance (Rs) were determined within this range. A considerable decrease in Rs value was recorded for 5CBCY tape compared to 5CBCY sintered samples. A series of symmetrical cells with 5CBCY electrolytes was analysed. The lowest ASR values for the investigated cells were found for a newly elaborated LSCF–5CBCY cathode as well as for a Ni–5CBCY anode. These electrode materials appear to be suitable for 5CBCY-electrolyte-supported solid oxide fuel cells.
KeywordsSolid oxide fuel cells Ceramic tape Ceramic proton conductor BaCe0.9Y0.1O3
Since Iwahara’s 1980 discovery of the high-temperature ceramic proton conductivity of BaCeO3-based materials, there has been growing interest in certain perovskite oxides, based on their potential application as proton-conducting ceramic membranes in electrochemical devices such as solid oxide fuel cells, solid oxide electrolysers, reactors for the oxidative dehydrogenation of certain organic compounds, and gas sensors for monitoring hydrogen in gas atmospheres [1, 2, 3, 4]. In general, the doped cerates MCe1−xYxO3−δ, M = Ba2+or Sr2+, exhibited higher electrical conductivity but limited chemical stability in CO2 or H2O gas atmospheres when compared, respectively, to the zirconates MZr1−xYxO3, Ba2+and Sr2+, where: 0 < x < 0.2 [5, 6, 7]. Nowadays, research in the field of SOFCs is focused mainly on an intermediate temperature (IT) range of 500–750 °C. In SOFCs with oxide electrolytes, water is produced on the anode side, which dilutes the fuel. Ceramic proton conductors based on BaCe1−xYxO3, 0 < x < 0.3, appear to be valuable components for developing cost-effective SOFCs operating within an IT range of 500–700 °C. Protons are characterised by greater mobility than oxygen ions because of their much smaller size. The diffusion of protons is also much less temperature dependent than that of oxygen ions. Therefore, at lower temperatures, it is potentially easier to obtain greater conductivity for proton conductors than for oxygen ion conductors. This provides a strong incentive (as well as the potential) to lower electrical resistivity and to improve chemical stability in gas atmospheres involving CO2 or H2O [8, 9]. Previously, it had been found that the partial substitution of barium for calcium in (Ba1−xCax)(M0.9Y0.1)O3 solid solutions, where M = Ce, Zr and 0 < x< 0.1, led to an improvement in electrical conductivity compared to initial BaM0.9Y0.1O3, M = Ce, Zr, samples .
Some potential strategic actions were elaborated to improve the physicochemical properties of BaCeO3-based materials [11, 12, 13]. It was found that the introduction of strontium into Ba1−xSrxCe0.9Y0.1O3, where 0 < x < 0.2, led to a small decrease in ionic conductivity compared to the initial BaCe0.9Y0.1O3 material, but improved chemical stability in gas atmospheres involving CO2 or water [14, 15]. An elaborated method involving the fabrication of a thin layer of proton-conducting ceramic appears to be one possible way to lower electrical resistance. Tape casting is a low-cost process particularly well suited for the fabrication of thin (100–1000 μm) flat components for solid oxide fuel cells, enabling the production of a wide variety of controlled morphologies, from highly porous to fully dense microstructures such as electrodes and electrolytes [16, 17, 18, 19]. It is widely known that a planar electrolyte-supported SOFC constitutes an attractive geometry for an SOFC stack configuration for which relatively large and flat cell components are required [20, 21].
Costa et al.  investigated the potential for forming BaCe0.9Y0.1O3 ceramic tapes using aqueous or non-aqueous slurries. It was found that a water-based system was unsuitable for manufacturing ceramic tapes involving barium cerate due to a strong tendency towards rapid hydrolysis. On the other hand, the application of a tape-casting system involving an organic medium enabled the fabrication of a BaCe0.9Y0.1O3 electrolyte as well as NiO–BaCe0.9Y0.1O3 anode-based materials. The analysis of the results is limited to X-ray diffraction analysis and SEM observation of the obtained materials. The literature is lacking in data reflecting the analysis of physicochemical properties of BaCeO3-based ceramic tape crucial for application as an electrolyte support in the construction of an intermediate temperature solid oxide fuel cell.
The aim of this study was to elaborate an organic tape-casting method for the fabrication of ceramic tape involving a Ba0.95Ca0.05Ce0.9Y0.1O3 membrane as a component of an electrolyte-supported SOFC. Special attention was paid to: (1) determining the physicochemical properties of BaCeO3-based tape; (2) elaborating cathode and anode composite materials suitable for the construction of a Ba0.95Ca0.05Ce0.9Y0.1O3-tape electrolyte-supported solid oxide fuel cell which could be operated in the temperature range 500–750 °C.
Ba0.95Ca0.05Ce0.9Y0.1O3 powder preparation
Ba0.95Ca0.05Ce0.9Y0.1O3 (hereafter: 5CBCY) monophase powder was synthesised by means of a solid-state reaction. The starting reagents used to synthesise 5CBCY powder were barium carbonate and calcium carbonate (Avantor Performance Material Poland S.A, p.a.), yttrium(III) oxide (Sigma-Aldrich, 99.99%), and cerium oxide (Acros Organics, 99.9%). Stoichiometric amounts corresponding to the formula Ba0.95Ca0.05Ce0.9Y0.1O3 were homogenised and milled in a planetary ball mill (Retsch PM 100). A small portion of mixed reagents corresponding to the formula Ba0.95Ca0.05Ce0.9Y0.1O3 was calcined within a temperature range of 900–1150 °C for 2 h. Following calcination, analysis of phase composition was performed using X-ray diffraction analysis. The monophase of 5CBCY was found at a temperature of 1150 °C. It was decided to synthesise a 200-g portion of powder at 1200 °C for 2 h. The 5CBCY powder was ground in dry ethanol with zirconia grinding media.
Preparation of slurry and 5CBCY ceramic tapes
The organic medium used for tape casting included polyvinyl butyral (Kuraray Europe GmbH) used as a binder, a plasticiser based on carboxylic acid esters (Zschimmer & Schwarz), and a mixture of ethanol (Avantor Performance Material Poland S.A, p.a.) and methyl ethyl ketone (Chempur, Poland) used as a solvent. The 5CBCY powder was mixed in a polythene container with an organic medium at a mass ratio of 54:46, attaining the form of slurry, which was then mixed and ground by means of ball milling to ensure homogenisation and destruction of agglomerates. The grinding process proceeded for 24 h, using high-speed rollers. The resulting slurry underwent deaeration for 24 h at a reduced rotation speed and subsequently for 2 min in low-pressure conditions. Then the slurry was transferred to a tape-casting device. The gap between the ‘doctor blade’ (cast squeegee) and tape-casting table surface amounted to 0.8 mm. Green tape was cast at the rate of 2 cm s−1 on a PTFE surface covered with anti-adhesive coating. Subsequently, the tape was dried in an air-conditioned room for 24 h. After the drying process, the thickness of the tape was 0.28 mm. At this stage, 28-mm disc-shaped samples were laser-cut from the tape. The 5CBCY samples were sintered within a temperature range of 1450–1600 °C. The 5CBCY ceramic tapes characterised by the highest relative density were selected for further investigations.
Electrochemical investigations of 5CBCY ceramic electrolytes in symmetrical solid oxide cells and in hydrogen–oxygen solid oxide fuel cells
H2, Pt|5CBCY|Pt, O2
O2, LSCF|5CBCY|LSCF, O2
O2, LSCF–GDC|5CBCY|LSCF–GDC, O2
O2, LSCF–5CBCY|5CBCY|LSCF–5CBCY, O2
H2, Ni–5CBCY|5CBCY|Ni–5CBCY, H2
In the case of cell (1), the 5CBCY electrolyte was manufactured in the shape of discs (diameter ~ 20 mm), but differences in thickness were applied in this study: 5CBCY ceramic tape with a thickness of 0.14 mm (hereafter: 5CBCY-T), and disc-shaped sintered samples with a thickness of 1.4 mm (hereafter: 5CBCY-S). Porous platinum electrodes (the active surface of each electrode equalled about 1.3 cm2) were applied in the initial electrochemical tests described in this paper. The electrodes were screen-printed from commercial Pt paste. Electrolytes with Pt electrodes were heated at 1000 °C for 1 h to obtain porous forms for both electrodes.
(La0.60Sr0.40)0.95Co0.20Fe0.80O3−δ (LSCF) was chosen from the group of oxide cathode materials for this investigation of the potential of prepared optimised cathode materials for an electrolyte-supported IT-SOFC with a 5CBCY electrolyte. Although LSCF is known to be a good cathode material for an IT-SOFC with a ceramic oxide electrolyte, some authors have reported that it also seems to be a promising cathode for SOFCs with a ceramic proton-conducting membrane [23, 24]. The electrochemical behaviour of an LSCF|5CBCY interface under prolonged cathodic polarisation was investigated. Two kinds of LSCF electrode were applied in this study. In the first case, a quasi-point electrode was made from dense sintered LSCF. The LSCF samples were prepared from commercial powders supplied by Fuel Cell Materials, USA. The pellets were isostatically pressed under 250 MPa and sintered at a temperature of 1200 °C for 24 h. LSCF cathodes were formed from previously sintered samples in the shape of pyramids by means of a circular saw and diamond files. LSCF quasi-point electrodes were placed along with a 5CBCY ceramic electrolyte in a custom-designed, handmade electrochemical setup [25, 26]. Only the end of the pyramid pointed at the cathode was placed in contact with the electrolyte. Electrochemical measurements of an LSCF|5CBCY system under prolonged polarisation, under negative potential from − 0.05 to − 0.7 V, were performed in air at 700 °C for 100 h. A porous monophase LSCF cathode and an LSFC-GDC cathode (thickness ~ 35 μm) were manufactured using the screen printing method. Commercial LSCF or LSCF–GDC paste was screen-printed on both sides of the electrolyte sample and then sintered at 1200 °C for 2 h with a heating and cooling rate of 1 °C min−1. A new composite LSCF–5CBCY cathode was also elaborated and investigated in this study. LSCF–5CBCY cathode composite paste was obtained by mixing equal equilibrium amounts of LSCF and 5CBCY powders in an agate mortar and then adding the organic carrier. The prepared paste was applied via screen printing to the 5CBCY electrolyte surface. An attempt to elaborate new Ni–5CBCY anode material for a 5CBCY electrolyte-supported IT-SOFC was also undertaken. Ni–5CBCY anode composite paste was prepared by mixing 5CBCY electrolyte powder with nickel oxide powder (Acros Organics) in an agate mortar. Subsequently, an organic carrier was added and ground to a homogeneous consistency. The prepared paste was applied by screen printing to the 5CBCY electrolyte surface. The anode was heated at 1400 °C for 4 h in air. In this way, a series of symmetrical solid oxide fuel cells (2)–(5) with 5CBCY tape were obtained.
Symmetrical button solid oxide fuel cells (2)–(5) with porous LSCF, LSCF–GDC, and LSCF–5CBCY cathodes were used to determine polarisation resistance Rp and area specific resistance ASR within the temperature range 550–700 °C.
Electrochemical measurements were performed using a Solartron SI 1287 Electrochemical Interface with a 1255B Frequency Response Analyzer. Measurements were performed in dry and wet air for cell (2) with LSCF electrodes. The amplitude of the sinusoidal voltage signal was 10 mV. The Minuit [27, 28] program was used to fit the parameters of the equation describing the assumed equivalent electrical circuit (EEC).
Analytical methods of evaluation of the physicochemical properties of a 5CBCY electrolyte prepared in the form of powder and as a ceramic tape
XRD measurements with a PANalytical Empyrean system employing monochromatic CuKα radiation was used to determine the phase composition to determine the phase composition of 5CBCY applied powder, sintered ceramic tape, an LSCF–C5BCY composite cathode, and Ni–5CBCY anode material. The lattice parameters of all investigated materials were determined using the Rietveld method. The morphology of 5CBCY particles was then observed using scanning electron microscopy (SEM). Ultra-high-resolution scanning electron microscopy with a Nova NanoSEM 200 (FEI Europe) was used for all SEM observations.
The particle size distribution of ground 5CBCY powder was determined in an ethanol suspension with a Mastersizer 2000 (Malvern Instruments) laser particle size analyser. The surface area of the powder was determined using the isotherm BET method (Nova 1200e, Quantachrome Instruments). The powder was subsequently applied to the preparation of slurry for the tape-casting process.
Thermal analysis (TGA-DTA, STA 449 F3 Jupiter thermal analyser, NETZSCH, USA) was used to determine the thermal effect occurring during the heating of green tape prepared from 5CBCY within a temperature range of 20–1000 °C. The measurements were performed in air; a ramp of 10 °C min−1 was applied. Confocal microscopy (Olympus LEXT OLS4000) was used to observe the surface of the 5CBCY tape before sintering. Dilatometry was used to determine the variations in the dimensions of samples heated within a temperature range of 25–1600 °C. The samples were measured using a NETZSCH model DIL 402. The density of 5CBCY samples was determined using the Archimedes’ method. Scanning electron microscopy was used to observe the microstructure of 5CBCY tape sintered in air in a temperature range of 1450–1600 °C for 2 h. SEM observation was also carried out for the microstructure of the obtained LSCF, LSCF–5CBCY, and Ni–5CBCY electrodes as well as for a cross section of the 5CBCY sintered ceramic tape and cross sections of all investigated SOFC cells.
The dependence of particle sizes (Fig. 2a) on volume indicated that this distribution is rather narrow: 80 vol% of all 5CBCY particles range in size from 2.42 to 20.99 μm. The parameters D10 (2.42 μm), D50 (5.64 μm), and D90 (20.92 μm), determined by means of particle size analysis, were also calculated. Analysis of SEM images (Fig. 2b) shows that 5CBCY powders are characterised by isometric particle sizes ranging from 0.5 to 20 μm.
The same form of agglomerates was observed using SEM. The data obtained from laser measurements are in close agreement with the SEM analysis. The isometric shape of 5CBCY particles and their narrow size distribution are favourable for the preparation of slurry for casting.
During the sintering process, 5CBCY green tape was subjected to the irreversible variation that occurs with an increase from room temperature to the final heating temperature. Knowledge of the thermal effects which occur during temperature increases is necessary for the elaboration of suitable sintering conditions in order to release the organic medium from the ceramic part, to reduce porosity, and to obtain gas-tight samples without cracks or certain defects.
A ceramic electrolyte designed for planar SOFC application should be contained in flat and dense tape with a homogenous microstructure. Electrolytes for SOFC application require a relative density above 95 wt%. The presence of porosity leads to a decrease in ionic conductivity and the mechanical properties of ceramic shapes.
All measurements were performed in analogous conditions. OCV values calculated using the Nernst equation were also included in the graph. Analysis of these values within the temperature range 450–750 °C indicated a correlation between a slight decrease in OCV values and increasing temperatures. Within a temperature range of 500–600 °C, the OCV of cell (1) reached values slightly higher than 1 V. The measured values of OCV from the Pt|5CBCY|Pt cell within the temperature range 500–600 °C in series (1) and (2) were very close to 1 V. A decrease in OCV values was observed within a higher temperature range, i.e. 650–750 °C. In other research, Z. Sun et al. also recorded a decrease of OCV values for a hydrogen–oxygen SOFC with BaZr0.9Y0.1O3 or BaZr0.9Y0.1O3 with a calcia addition in the temperature range 500–700 °C. In the case of Pt|BaZr0.9Y0.1O3|Pt or Pt|BaZr0.9Y0.1O3 + 4% mol CaO|Pt SOFC cells, a marked tendency towards a decrease in OCV values was also observed at temperatures higher than 600 °C. An increase in electron–hole conduction in BaZrO3-based electrolytes in an air atmosphere at higher temperatures was pointed out as the main reason for the reduced OCVs in the investigated cells. In the case of the Pt|BaZr0.9Y0.1O3 + 4%mol CaO|Pt cell, an OCV value of about 0.93 V was recorded at 700 °C, compared to 0.99 V at 500 °C [33, 34]. Very similar values of OCV were reached for a Pt|5CBCY|Pt cell at the above-mentioned temperatures. In the case of BaCeO3-based materials, the OCV drop may result as well from a decrease in the ionic transport number at higher conductivity due to a possible increase in partial electronic conductivity.
Based on the initial electrochemical investigations of an IT-SOFC involving a 5CBCY electrolyte, the high values of OCV obtained for cell (1) also confirmed the gas-tightness of the electrolyte manufactured in the form of ceramic tape as well as in that of a pressed and sintered sample.
One of the strategic goals leading to the improved performance of a SOFC with a 5CBCY electrolyte should be the identification of a suitable cathode material. The oxygen reduction reaction (ORR) occurring at the cathode of a SOFC causes most of the energy losses, leading in turn to a reduction in the cell’s energy efficiency. In the case of a ceramic fuel cell with a proton ceramic membrane, water is formed on the cathode side, complicating the cathode reaction mechanism and resulting in a higher level of polarisation resistance in the cathode. Although significant efforts have been exerted towards developing suitable cathodes for ceramic fuel cells with proton-conducting membranes (PCFCs), the problem of selecting an optimum cathode material has not yet been solved. Information describing the electrochemical behaviour of an LSCF cathode under long-term cathodic polarisation is lacking. In this study, quasi-point electrodes as well as classical porous LSFC electrodes were applied.
The electrochemical performance of the cathode follows the order LSCF–5CBCY > LSCF > LSCF–GDC. The enhanced performance of the LSCF–5CBCY composite cathode compared to the LSCF cathode can be ascribed mainly to the introduction of proton-/oxide-ion-conducting pathways. In the case of 5CBCY in dry air, oxide ion conductivity/electronic conductivity was observed. Composite cathodes involving MIECs, electrocatalysts, and oxygen ion conductors (O-composite cathodes) were also investigated for solid oxide fuel cells with proton-conducting membranes. He  found that the enhancement of electrochemical activity is the effect of addition of oxygen-ion conductor and could be attributed to the improvement on the surface exchange dynamics or providing the alternative oxygen diffusion path.
Some authors have reported that the electrochemical behaviour of LSCF- or SSC-infiltrated GDC O-composite electrodes in symmetrical cells exhibits lower polarisation resistance . In this study, the introduction of the oxide-ion conductor GDC to the LSCF electrode did not lead to an improvement in performance compared to a monophase LSCF electrode. The lowest ASR value in the investigated temperature range was observed at 550–700 °C for the LSCF–5CBCY electrode. A decrease in cathode polarisation resistance was observed due to an increase in ionic (oxide/proton) conductivity within a temperature range of 500–600 °C.
Figure 13 presents the ASR values determined for H2,Ni–5CBCY|5CBCY|Ni–5CBCY,H2 in the same temperature range. As can be seen, the elaborated Ni–5CBCY appears to be a suitable anode for an E-SOFC operating with a 5CBCY electrolyte. These results confirmed the actual potential for constructing a 5CBCY-electrolyte-supported SOFC with a LSCF–5CBCY cathode and a Ni–5CBCY anode.
Slurry for preparing a Ba0.95Ca0.05Ce0.9Y0.1O3 ceramic tape was successfully obtained from monophase ground 5CBCY powder originating from a solid-state reaction, in combination with organic binders. Flat gas-tight 5CBCY ceramic tape was obtained after sintering at 1600 °C for 2 h in air. The uniform microstructure of the tape, with grains 1–10 μm, was observed with the use of SEM. The initial electrochemical tests performed on solid button oxide fuel cells confirmed the high quality of the prepared ceramic tape from the 5CBCY electrolytic material. The superiority of an IT-SOFC with 5CBCY ceramic tape, as opposed to 5CBCY disc samples, was proven. The ASR values obtained for a newly elaborated LSCF–5CBCY composite cathode are compatible with its application as a component of a 5CBCY electrolyte-supported solid oxide fuel cell. A Ni–5CBCY composite anode with lower ASR resistance was also elaborated.
The paper was completed under a contract with the AGH University of Science and Technology, Cracow, Poland (No. 18.104.22.1685). Some measurements were performed using scientific equipment belonging to the laboratories of the AGH-UST Energy Centre, Cracow, Poland.
- 24.Ii H, Choi S, Yoon KC, Son H, Kim BK, Lee H, Lee J. Influence of wet atmosphere on electrical and transport properties of lanthanum strontium cobalt ferrite cathode materials for protonic ceramic fuel cells. Solid State Ionics. 2013;249:249–50.Google Scholar
- 25.Raźniak A, Tomczyk P. Application of microelectrodes for investigation of the oxygen electrode reaction in selected electrolytes. Mater Sci Poland. 2008;26:195–206.Google Scholar
- 32.Ding Y, Li Y, Huang W. The role of Ba concentration on the structural characteristics and electrical conductivities of BaxCe0.9Y0.1O3−α. Mater Res Bull. 2017;205:122–8.Google Scholar