A zero-dimensional model was implemented to simulate the performance of a solid oxide fuel cell (SOFC) with Mo-doped ceria (CMO) anode using humidified (3%v/v water) hydrogen or syngas (50%v/v H2 – 50%v/v CO) as fuel. The model was validated preliminarily through experimental data generated at 800ºC using a SOFC with coated electrodes supported in the electrolyte (a 20 mm diameter yttria-stabilized zirconia, YSZ, disc) with a 14 mm diameter CMO|CMO-YSZ anode and a 14 mm diameter LSM-YSZ|LSM (lanthanum strontium manganite) cathode. Polarization (cell voltage vs current density, ΔEcell vs j) and power density (power density vs current density, Pelec vs j) curves were generated via linear sweep voltammetry connecting the SOFC to a Gamry Reference 3000 potentiostat/galvanostat/ZRA using platinum mesh and wires.The open circuit voltage (OCV) of the SOFC operating with humidified hydrogen was modelled directly using Nernst equations for the electrochemical hydrogen oxidation and oxygen reduction reactions. For the case of the SOFC operating with humidified syngas, the electrochemical hydrogen and carbon monoxide oxidation reactions occur simultaneously at the anode. Hence, the Nernst equation was corrected by each anodic reaction selectivity determined solving mass balances coupled to an electrochemical kinetics model.Figure 1.a shows the fitted polarization and power density curves. Maximum SOFC power densities of ca. 1295 W m-2 (at 0.53 V and 2434 A m-2) and ca. 1084 W m-2 (at 0.52 V and 2106 A m-2) are predicted for the cell operating with humidified hydrogen and humidified syngas as fuel, respectively. These values are 2 and 5% lower than the experimental values. The OCVs calculated are ca. 1.10 V and 1.01 V, respectively, which are 5 and 8% higher/lower than the experimental values. The differences observed are attributed to the non-ideality of the gases due to their interaction with the electrode surfaces, and the possible infiltration of air into the anodic chamber [1,2].The SOFC ohmic losses calculated represent ca. 72% of the total voltage losses when operating at the maximum power density, which are attributed to the low electrical conductivities of the CMO|CMO-YSZ anode (CMO electrical conductivity is ca. 0.501 S m-1 at 800ºC [3]) and the thick electrolyte (ca. 2.203 S m-1 at 800ºC [4]). Improvements are being introduced to the anode material by including nickel, which will contribute to increase the hydrogen oxidation kinetics while CMO protects the electrode from carbon deposition when syngas is used as fuel. Presently, activation losses represent ca. 28% of the total voltage losses at the maximum power density. It has been reported that for SOFCs using syngas as fuel, the anode active sites are more occupied by carbon monoxide molecules due to their slower oxidation kinetics in comparison with hydrogen. Nevertheless, carbon monoxide is also susceptible to undergo a surface reforming process via the water-gas shift reaction. Such phenomenon becomes more relevant at high carbon monoxide concentrations, which was the case in this study (50% v/v CO), thus favoring the electricity generation through hydrogen oxidation at the anode surface [2].For the SOFC using humidified syngas as fuel, the model assumes that the desorption equilibrium pressure of hydrogen (pH2 *) on the surface of molybdenum in the CMO|CMO-YSZ anode is equal to that reported for humidified hydrogen on the surface of nickel in Ni-YSZ anodes (0.09 atm) [5]. Figure 1.b shows the results obtained for the sensitivity analysis of the model as a function of pH2 *, which indicates that unless its value is increased by ca. 220% over 0.09 atm the exchange current density for carbon monoxide oxidation (j0,CO) does not go below its lower confidence limit. It is unlikely that this parameter reaches such level of discrepancy because it implies that hydrogen adsorption at the anode surface is strongly diminished, thus electrochemical oxidation kinetics decrease considerably which is not consistent with the experimental observations.Future work includes the analysis of the kinetic parameters dependence on the temperature and the fuels partial pressures.
Read full abstract