Abstract
Solid Oxide Cells (SOCs) are under intensive development due to their great potential to meet the 2030 targets for decarbonization. One of their advantages is that they can work in reversible mode. However, in respect to durability, there are still some technical challenges. Although the quick development of experimental and modeling approaches gives insight into degradation mechanisms, an obligatory step that cannot be avoided is the performance of long-term tests. Taking into account the target for a commercial lifetime is 80,000 h, experiments lasting years are not acceptable for market needs. This work aims to develop accelerated stress tests (ASTs) for SOCs by the artificial aging of the fuel electrode via redox cycling, which follows the degradation processes of calendar aging (Ni coarsening and migration). However, it can cause irreversible damage by the formation of cracks at the interface anode/electrolyte. The advantages of the developed procedure are that it offers a mild level of oxidation, which can be governed and regulated by the direct impedance monitoring of the Ni network resistance changes during oxidation/reduction on a bare anode sample. Once the redox cycling conditions are fixed and the anode/electrolyte sample is checked for cracks, the procedure is introduced for the AST in full-cell configuration. The developed methodology is evaluated by a comparative analysis of current voltage and impedance measurements of pristine, artificially aged, and calendar-aged button cells, combined with microstructural characterization of their anodes. It can be applied in both fuel cell and electrolyzer mode. The results obtained in this study from the electrochemical tests show that the artificially aged experimental cell corresponds to at least 3500 h of nominal operation. The number of hours is much bigger in respect to the microstructural aging of the anode. Taking into consideration that the duration of the performed 20 redox cycles is about 50 to 60 working hours, the acceleration factor in respect to experimental timing is estimated to be higher than 60, without any damaging of the sample. This result shows that the selected approach is very promising for a large decrease in testing times for SOCs.