Structure Property relationships in Solid Oxide Fuel Cells

Department of Ceramic Engineering University of Missouri - Rolla 222 McNutt Hall Rolla, MO 65401


The electrode reactions are a major cause of the energy losses in SOFC's, and limit their use to higher temperatures, typically 800-l000oC. The electrode reactions have received much attention aimed at better understanding the electrode kinetics and mechanisms, but are still very primitive in their basic understanding. The electrode microstructure and its corresponding reactivity has commonly been studied by DC and AC impedance techniques. A common method of examining electrode reactions employs surface-mounted reference electrodes, although this technique often limits the experiment to examination of one electrode. In this study a new technique has been developed of utilizing a Pt voltage probe placed internally into the electrolyte to measure the I-V and impedance spectra of both electrodes operating under cell conditions. Unlike surface mounted electrodes which need to be concerned with distance and dimensions of reference electrodes with respect to working and counter electrodes the internal Pt voltage probe is centered internally at a known depth within the electrolyte and between corresponding electrodes.

The internal Pt voltage probe has been used in this research program to investigate the microstructure o property relations in solid oxide fuel cells (SOFC's) in order to better understand the mechanisms involved in cell performance. The aim is to fabricate SOFC's with controlled microstructures utilizing La^Sr^MnOj (LSM), yttria stabilized zirconia (YSZ), and Ni-YSZ composites as the cathode, electrolyte, and anode, respectively. Ideally, the electrode materials would be tailored for an increased reaction rate (grain size < 1 pm), be stable with time (> 10,000 h), have a thermal expansion match to YSZ (a=l 1 x 10"6/°C), show limited chemical interaction with the electrolyte, and show no degradation in electrical performance. This paper describes just a few of the starting powder characteristics, electrical conductivity and overpotential measurements, and resultant microstructures as a function of processing conditions (i.e. powder calcination temperature, and annealing temperature) and composition for the electrolyte and cathodes.


Sinple Cell Fabrication

The YSZ electrolyte used in this investigation was self supporting, and the cathodes were applied via screen printing onto a pre-sintered dense electrolyte. The commercially-available YSZ (Zirconia Sales of America Inc.) is a fully stabilized (8 mole % Y2O3), co-precipitated powder. This particular powder was chosen because of its low cost, ~$70/kg, low impurity content and low densification temperature, -1400°C. The YSZ powder had a primary particle size of approximately 250 nm and a corresponding BET surface area of - 8.0 m2/g. YSZ powders were initially dried at i 50°C to remove any physically bonded water and then mixed with a commercially available binder system from Ferro Corp, B73210. The slip was ball-milled with Zr02 media for 24 - 48 h until the powder was well dispersed. The slurry was tape cast; green thicknesses were - 50-75 pm. Circular samples of 2 inch diameter were cut out of the tape and a Pt voltage probe (38 mm long x 0.3 mm wide) was screen printed onto the YSZ tape. Six tapes were stacked and laminated at 3000 psi for 10 min at 70°C in such a way as to place the Pt probe in the center of the fired specimen. The thermal processing schedule for binder removal and densification of the YSZ laminates was 0.5°C/min to 350°C, hold for 1 h, 3°C /min up to 1450°C, hold for 2 h, and then cool at 3°C/min. Sintered electrolytes were 3.2 cm in diameter, and 200 pm thick.

LaxSro.2Mn03 (0.7 < x < 0.79) compositions were synthesized by the liquid mix technique using La2(C03)3, SrC03, and MnCO_, as standardized raw materials. The carbonates were dissolved in HNO„ ethylene glycol, citric acid and water in glass beakers. Heating resulted in the formation of a polymeric precursor solution which was then oxidized at ~300°C. The powder was vibratory milled dry for 4 h with Zr02 media, then calcined in MgO crucibles at temperatures of 800 - 1200°C in increments of I00°C. Soak times for all calcinations was 4 h. The powders were vibratory milled again using the same conditions as described before. Powder crystallinity, phase, and surface area were characterized using X-ray diffraction and BET techniques as a function of calcination temperature. The electrode powders were mixed with a commercial resin solution,

BX018-16, from Ferro Corp using a three roll mill. The cathode compositions were screen printed onto dense YSZ electrolytes and sintered at temperatures between 1200°C and 1400°C in 100°C increments for a 1 h hold, with a heating and cooling rates of 3°C/min. A primary goal of this investigation was to vary the grain size and porosity of electrode microstructures and the their impact on electrode performance, therefore powders were calcined and sintered at various temperatures. Cathodes were porous, exhibited grain sizes on the order of 1pm, and gave resultant dimensions of 0.635cm x 0.635cm and - 20pm thick.

A Pt grid (0.2 mm line width and 0.2 mm spacing between lines) was screen printed on the electrodes for cell performance experiments to act as a current collector but also to allow gas diffusion to the electrode/electrolyte interface.

Electrical Characterization

Electrical characterization of single cells utilizing the internal Pt voltage probe was investigated to simultaneously separate the losses attributed to each component (anode, electrolyte, cathode) and their interfaces (cathode/electrolyte and anode/electrolyte) during cell operation. The , cell performance studies were focused on the reaction "kinetics^ at the interfaces whereas DC conductivity measurements were performed to investigate the resistive losses of each component as a function of time, composition, and preparation condition. Electrochemical (I-V) measurements , were carried out using a five electrode configuration which allowed for separation of anode and cathode overpotentials during operation. I-V behavior was performed on both half cells and complete cells using an Anatronics Current/Voltage Control Fuel Cell Testing Module, a Keithley Model 196 Microvolt Meter, and a Fluke 27/FM Multimeter. The Fuel Cell Testing Module was placed in the voltage control mode thus enabling the desired cell voltage and corresponding current to be measured

0 0

Post a comment