A zero-dimensional and isothermal stationary model demonstrably predicting the current-voltage (C/V) characteristic of an anode supported SOFC single cell is for the first time verified for C/V characteristics measured in electrolysis mode. The accuracy of the presented model was increased by including the actual cell temperature under current load, determined by an impedance-based temperature measurement routine. C/V characteristics measured at 800°C in the range from 0.66 V to 1.6 V for H2O:H2 compositions 70:30 and 30:70 reveal a pronounced asymmetric operation of the fuel electrode supported cell in electrolysis mode. This experimentally observed behavior is accurately reproduced by the model and is explained by (i) increasing polarization losses related to Knudsen diffusion and (ii) decreasing reaction rate in dry conditions at the fuel electrode at high current densities in electrolysis mode.

A zero-dimensional and isothermal stationary model demonstrably predicting the current-voltage (C/V) characteristics of anode supported Solid Oxide Fuel Cells is proven to accurately reproduce these C/V characteristics in reverse (Electrolyzer) operation mode as well. The asymmetry observed in the C/V curves of both modes is totally reproduced and also explicitly explained.

Electrochemical reaction kinetics at the electrodes of Solid Oxide Cells (SOCs) were investigated at 700°C for two cells with different fuel electrode microstructures as well as on a third cell with a reduced active electrode area. Three fuel mixtures were investigated – hydrogen/steam and model reformate fuels–hydrogen/carbon-dioxide and hydrogen/methane/steam. It was found that the electrode kinetics at the fuel electrode were exactly the same in both reformates. The hydrogen/steam fuel displayed 5–7% faster kinetics than the reformate fuels. 19% faster kinetics were recorded for the cell with a finer microstructure. The measured gas conversion impedance was compared with models in literature for both the 16- and the 2 cm2 cells. The continuously stirred tank reactor (CSTR) AC model approximated the overpotential of the smaller cells (2 cm2) with greater accuracy in the current density range 0–0.5 A/cm2 while the plug flow reactor (PFR) model although derived for the case of zero DC bias predicted the 16 cm2 cell ASR better than the zero bias CSTR model. Furthermore, the gas conversion impedance in the hydrogen/steam fuel split into two processes with opposing temperature behavior in the reformate fuels. By using a 87.5% smaller active electrode area the gas conversion impedance was diminished in the hydrogen/steam fuel at (the same absolute) high fuel flow rates. In both reformates, the second and third lowest frequency processes merged into a single process as the gas conversion was reduced. The SOC with finer electrode microstructure displayed improved kinetics.

Our zero-dimensional and isothermal stationary model predicts the current-voltage (C/V) characteristics of anode supported SOFC single cells in the direct as well as in the reverse (electrolyzer) operation mode. An extended model is presented for the electrolyzer mode, taking the corrected cell temperature as additional variable into consideration. This cell temperature is derived from measurement applying single-frequency electrochemical impedance spectroscopy (SF-EIS).

AC and DC techniques were applied to investigate the electrochemical reaction kinetics of porous composite Ni/8-mol% yttria-stabilized zirconia (Ni/8YSZ) solid oxide cell (SOC) electrodes using a novel pseudo-3-electrode cell geometry. From OCV impedance spectra an activation energy Ea of 1.13 eV, pre-factor γan of 3.7∙105∙T, hydrogen and steam partial pressure dependencies a and b respectively of -0.07 and 0.22 were determined. DC current density vs. overpotential curves compared with those predicted using the determined kinetic parameters. Apparent Butler-Volmer charge transfer coefficients α were determined from the current density vs. overpotential curves. Values ranging from 0.57 at 650 °C to 0.64 at 850 °C were determined from the anodic branch and 0.85 to 0.81 from the cathodic branch in the same range, with higher fitting accuracy in the anodic branch. The lower fitting accuracy of the cathodic branch and the need for different α values for each branch suggests that a simple BV model of the measured electrode kinetics is insufficient and/or different reaction mechanisms might be occurring in anodic vs cathodic polarization.