Keith L. Duncan

Determination of Surface Exchange Coefficients of LSM, LSCF, YSZ, GDC Constituent Materials in Composite SOFC Cathodes

A novel approach to processing and modeling isothermal isotope exchange (IIE) data was developed to extract kinetic rate coefficients for the oxygen reduction reaction (ORR) on cathode materials used for solid oxide fuel cells (SOFC). IIE is capable of testing powders with particle sizes on the nano-scale, where the effective sample thickness (particle size) was shown to be below the characteristic thickness (L c ) for ionically conducting materials. This allows for accurate kinetic measurements in surface exchange controlled regimes, in contrast to secondary ion mass spectrometry depth profiling and electrical conductivity relaxation techniques where sample thicknesses are typically in the diffusion limited or mixed regimes. Surface exchange coefficients (k * ) were extracted and L c values calculated from cathode (La 0.6 Sr 0.4 )(Co 0.2 Fe 0.8 )O 3-δ (LSCF) and (La 0.8 Sr 0.2 )MnO 3 (LSM), and electrolyte (Ce 0.9 Gd 0.1 )O 1-δ (GDC) and (Zr 0.8 Y 0.2 )O 2 (YSZ) materials. Additionally, diffusion coefficients (D * ) were extracted for LSM. In a surface exchange controlled regime LSCF exhibits a low activation energy for k * , while for LSM k * was observed to increase with decreasing temperature consistent with a precursor-mediated mechanism in which there is a negative apparent activation energy for the dissociative chemisorption of O 2 . GDC was shown to exhibit a low activation energy for k * , lower than YSZ, which is attributed to higher concentration of electrons in GDC than YSZ.

Surface and Bulk Defect Equilibria in Strontium-Doped Lanthanum Cobalt Iron Oxide

Strontium-doped lanthanum cobalt iron oxide is used in solid oxide fuel cells, oxygen separation membranes, and fuel reforming applications because of its mixed ionic and electronic conductivity. With changes in temperature and oxygen partial pressure (Po 2 ), the defect concentration changes and there is a corresponding change in strain (chemical expansion) and performance. In this paper, the oxygen nonstoichiometry of 0.5% A-site cation deficient La 0 . 6 Sr 0.4 Co 0.2 Fe 0 . 8 O 3-δ is studied in the temperature range of 600-900°C and modeled using a metallic model and three semiconductor models assuming itinerant electrons and holes or electrons localized on B-site, cations, with holes localized on either B-site cations or oxygen anions. The model consisting of holes localized on B-site cations fits the data best. The surface oxygen nonstoichiometry was determined by measuring oxygen loss of samples with different surface areas, and the surface is much more defective than the bulk. A long-term mass change measurement was also performed at 800°C in air, and a mass loss was observed that may be due to cation segregation to the surface.

The role of point defects in the physical properties of nonstoichiometric ceria

Explicit, analytical expressions have been derived for the dependence of electrical conductivity, chemical expansion, and elastic modulus on point defect concentration and oxygen partial pressure using a consistent approach. In developing the model, expressions were first derived for the functional dependence of defect concentration on the oxygen partial pressure for fluorite oxides. Expressions for the chemical expansion, elastic modulus, and electrical conductivity as functions of defect concentration are then derived and verified with experimental data for ceria (CeO2−δ) with consistently good fits. The same values for the material constants were used in all of the fits, further validating our approach.

Continuum-Level Analytical Model for Solid Oxide Fuel Cells with Mixed Conducting Electrolytes

A continuum-level analytical model was developed for the performance of solid oxide fuel cells (SOFCs) with mixed conducting electrolytes. The model was derived by coupling the Nernst-Planck equation for mass transport in the bulk with the Butler-Volmer equation for mass transport across an interface and by employing defect thermodynamics for boundary values. This approach allowed the usual assumptions of reversible electrodes and linear potential gradients (across the electrolyte) to be removed, thereby resulting in boundary defect concentrations that depend on potential (SOFC operating conditions) and allowing non-ohmic responses in mixed conducting electrolytes. Using only three fitting parameters, the authors validated the model through successful fits to experimental data from SOFCs with samaria- and gadolinia-doped ceria electrolytes, which are presently the mixed conducting electrolytes of greatest interest. The fits also allowed valuable information to be extracted from the experimental data regarding the overpotentials in the SOFC. Significantly, the anode, cathode, and electrolyte overpotentials were uniquely determined, each as a function of current. Hence, it was confirmed that performance losses due to electronic leakage through ceria electrolytes is confined to low load (low current and high voltage) conditions. Finally, the model was used to show that open-circuit voltage is a function of electrolyte thickness.

Surface Exchange Coefficients of Composite Cathode Materials Using In Situ Isothermal Isotope Exchange

Isothermal isotope exchange (IIE) was used to evaluate the surface kinetics of the oxygen reduction reaction on composite cathode materials. With IIE, powder materials can be tested to ensure surface exchange behavior is isolated. Macroscopic diffusion coefficients (D * ) and surface exchange coefficients (k * ) were extracted from 50:50 wt % composite cathode powders (La 0.6 Sr 0.4 )(Co 0.2 Fe 0.8 )O 3―δ /(Ce 0.9 Gd 0.1 )O 2―δ (LSCF/GDC), (La 0.8 Sr 0.2 )MnO 3 /(Y 2 O 3 ) 0.08 (ZrO 2 ) 0.92 (LSM/YSZ), and (La 0.8 Sr 0.2 )MnO 3 /(Ce 0.9 Gd 0.1 )O 2―δ (LSM/GDC) using IIE, and characteristic thicknesses (L c ) were calculated from the ratio of D * /k * . Mixed ionic and electronic conductor LSCF/GDC was shown to have the highest D * values between 500 and 850°C and k * values above 825°C. Adding an ionically conducting phase to LSM significantly improved the magnitude of k * , which below 825°C were highest for LSM/YSZ of the three composites. In addition, LSM/YSZ and LSM/GDC exhibited an apparent negative activation energy, which can be explained by a precursor-mediated mechanism for dissociative adsorption. Sample thicknesses (particle size) were shown to be well below characteristic thicknesses (L c ), ensuring that samples were in a surface exchange controlled regime and validating the accuracy of the extracted k * values.

Permeation Through SrCe0.9Eu0.1O3 − δ / Ni – SrCeO3 Tubular Hydrogen Separation Membranes

Thin-film (∼30 μm) SrCe 0.9 Eu 0.1 O 3―δ membranes on tubular porous Ni―SrCeO 3 supports were developed to investigate hydrogen permeability. Appreciable and stable hydrogen permeation through these thin-film membranes was observed under humid hydrogen conditions, and their [P H2 ] 1/4 dependence on hydrogen permeation flux agrees with Norby and Larrings model [Solid State Ion., 136/137, 139 (2000)] in which protons and electrons are the dominant charge carriers. The magnitude of the activation energy for hydrogen permeation flux suggests that the hydrogen permeation flux is limited by electronic conductivity. Hydrogen permeation fluxes of 7 cm 3 /min were achieved from 97% H 2 and 3% H 2 O at 900°C. On a membrane-area-normalized basis, this corresponds to a H 2 flux of 0.6 cm 3 /min cm 2 . The hydrogen permeation is stable in humid hydrogen but degrades in dry hydrogen due to SrCeO 3 decomposition in very low P O2 .

Effect of La2Zr2O7 on Interfacial Resistance in Solid Oxide Fuel Cells

The impact of La 2 Zr 2 O 7 (LZO) on interfacial resistance (R p ) at the La 0.78 Sr 0.20 MnO 3-δ /yttria-stabilized zirconia interface was studied upon isothermal sintering at 1200°C for 2-25 h. Quantification of triple phase boundary length was performed by applying focused ion beam/scanning electron microscopy (FIB/SEM) serial-sectioning techniques and classical stereology. Electrochemical impedance spectroscopy was used to characterize the R p . The effect of LZO formation on microstructural models for R p was analyzed with respect to previous works that did not include this effect. LZO formation modifies the TPB length, rapidly increases R p , and needs to be controlled in analytical microstructural R p models.

Determination of Electrochemical Performance and Thermo-Mechanical-Chemical Stability of SOFCs from Defect Modeling

The objectives of this project were to: provide fundamental relationships between SOFC performance and operating conditions and transient (time dependent) transport properties; extend models to thermo-mechanical stability, thermo-chemical stability, and multilayer structures; incorporate microstructural effects such as grain boundaries and grain-size distribution; experimentally verify models and devise strategies to obtain relevant material constants; and assemble software package for integration into SECA failure analysis models.

Oxygen Generation from Carbon Dioxide for Advanced Life Support

The partial electrochemical reduction of CO2 using ceramic oxygen generators (COGs) is well known and has been studied. Conventional COGs use yttria-stabilized zirconia (YSZ) electrolytes and operate at temperatures greater than 700 C (1, 2). Operating at a lower temperature has the advantage of reducing the mass of the ancillary components such as insulation. Moreover, complete reduction of metabolically produced CO2 (into carbon and oxygen) has the potential of reducing oxygen storage weight if the oxygen can be recovered. Recently, the University of Florida developed ceramic oxygen generators employing a bilayer electrolyte of gadolinia-doped ceria and erbia-stabilized bismuth oxide (ESB) for NASA s future exploration of Mars (3). The results showed that oxygen could be reliably produced from CO2 at temperatures as low as 400 C. These results indicate that this technology could be adapted to CO2 removal from a spacesuit and other applications in which CO2 removal is an issue. This strategy for CO2 removal in advanced life support systems employs a catalytic layer combined with a COG so that the CO2 is reduced completely to solid carbon and oxygen. First, to reduce the COG operating temperature, a thin, bilayer electrolyte was employed. Second, to promote full CO2 reduction while avoiding the problem of carbon deposition on the COG cathode, a catalytic carbon deposition layer was designed and the cathode utilized materials shown to be coke resistant. Third, a composite anode was used consisting of bismuth ruthenate (BRO) and ESB that has been shown to have high performance (4). The inset of figure 1 shows the conceptual design of the tubular COG and the rest of the figure shows schematically the test apparatus. Figure 2 shows the microstructure of a COG tube prior to testing. During testing, current is applied across the cell and initially CuO is reduced to copper metal by electrochemical pumping. Then the oxygen source becomes the CO/CO2. This presentation details the results of testing the COG.

A model for the spatial distribution and transport properties of defects in mixed ionic-electronic conductors part I: Defect concentration — Pressure relationships and the spatial distribution of defects

The spatial distribution and transport of defect species in a mixed ionic-electronic conductor (MIEC) depend on the activity of the external gas phase at either side of the MIEC, the defect equilibria and mobility of the species within the MIEC. Previous researchers [1, 2] developed models for defect distribution and transport based on various assumptions, the most prevalent of which is that the concentration of the ionic defect species is constant through the range of activities of the external gas phase in which the MIEC finds application. However, the defect equilibria for many MIECs do not support this assumption. It is desirable, therefore, to have explicit, analytical expressions for the spatial distribution of the defects as a function of position and the flux as a function of the activity gradient of the defect species and the gas phase. Towards the development of such a model, the defect concentration-oxygen partial pressure dependence in an oxide MIEC, such as Ce0.8Sm0.2O2−δ, is considered. From the defect equilibria, equations are developed for the dependence of the defect species on the oxygen activity in the oxide MIEC. Then, to relate the defect equilibria to the spatial parameters of the MIEC, an analytic expression for the variation of oxygen activity as a function of position in an electrolyte is derived. Equations relating oxygen activity to position and defect concentration to oxygen activity are then combined to obtain a defect concentration-position relation. All the relations are derived without the assumption of constant vacancy concentration. Finally, the results are compared with those of previous models.