No abstract

An electrochemical study of SOFC cathode degradation, due to poisoning by chromium oxide vapours, was performed applying 3‐electrode set‐ups. The cathode materials comprised LSM/YSZ and LSCF/CGO composites, whereas the electrolyte material was 8YSZ. The degradation of the cathode performance was investigated as a function of time under a current load of 0.2 or 0.4 A cm–2 and in the presence of Cr2O3 at 850 and 750 °C in air, dry or water saturated at room temperature, and compared to that of non‐Cr exposed reference specimens tested under, otherwise, the same conditions. This involved continuous logging of the DC current, and the voltage between the reference electrode and the working electrode in the 3‐electrode set‐up combined with frequent AC impedance measurements under current load which would allow the deduction of the cathode polarisation resistance (Rp). The duration of the tests ranged from 300 to 2,970 h. Both LSM/YSZ and LSCF/CGO cathodes were sensitive to chromium poisoning; LSCF/CGO cathodes to a lesser extent than LSM/YSZ. Humid air aggravated the degradation of the cathode performance. Post‐mortem electron microscopic investigations revealed several Cr‐containing compounds filling up the microstructure of the cathodes.

In this work, co‐electrolysis of steam and carbon dioxide was studied in a Topsoe Fuel Cell (TOFC®) 10‐cell stack, containing three different types of Ni/yttria stabilized zirconia (YSZ) electrode supported solid oxide electrolysis cells with a footprint of 12 × 12 cm. The stack was operated at 800 °C and –0.75 A cm–2 with 60% conversion for a period of 1,000 h. One type of the cells showed no long term degradation but actually activation during the entire electrolysis period, while the other two types degraded. The performance and durability of the different cell types is discussed with respect to cell material composition and microstructure. The results of this study show that long term electrolysis is feasible without notable degradation at 800 °C and a current density of –0.75 A cm–2.

Various models of the H 2 /H 2 O/Ni/SZ (SZ = stabilized zirconia) electrode kinetics have been presented in the literature in order to explain the reported experimental data. However, there has been a strong tendency of using a limited set of data to "verify" a given model, disregarding other data sets, which do not fit the model. We have inspected some models in the literature, and problems (e.g. no quantitative model has explained the large variation in reported values of apparent activation energy of the electrode kinetics) as well as strengths of the models are discussed. We point out experimental findings that a useful model must be able to explain such as difference in sensitivity to poisoning by H 2 S due to differences in the detailed composition of the SZ and large change in apparent activation energy by change in cermet preparation. Finally, we will point out some elements, which seem important for any realistic and useful mathematical model of the H 2 /H 2 O/Ni/SZ electrode.

No abstract