Iron or chromium based alloys are currently the most preferred materials for SOFC interconnects. The poisoning of the cathode of the membrane electrode assembly (MEA) due to evaporation of chromium species from the metallic interconnect (MIC) and the oxidation of the interconnect surface during stack operation at elevated (800-900 °C) temperatures are regarded to be the major mechanisms affecting the degradation behaviour of the SOFC stack. The oxidation at the cathode side of the interconnect can cause up to one third of the total stack degradation. Despite intensive investigations on corrosion stability of MIC for SOFC application over the past decades, the main attention was focused mainly on the material properties in the cathode side gas conditions (ambient air). The growth behavior and the composition of oxides grown in the SOFC anode gas environment as well as the influence of the oxides on the contact resistance between MIC and anode contacting are still not investigated comprehensively. The MIC materials have to be stable over more than 40,000 hours to satisfy a demand for a continuous increase of operation times of the SOFC stacks. Real time tests over such long time scales are cost intensive. A modification of standard material test procedures, with the aim to accelerate material degradation, is therefore necessary. In order to realize this, three ferritic steels (Crofer 22 APU, Crofer 22 H, AISI 441) and a chromium-based alloy were tested in SOFC anode gas environment in a temperature range between 725 – 875°C. The experiments were carried out under variation of the water vapor content in the gas mixture for different exposure times in order to create the accelerated degradation testing conditions for MIC and to investigate the behavior of these materials caused by the formation of growing chromium oxide based scales. Both gravimetric measurements and FESEM/EDX data, as well as polished microimage sections were analyzed to characterize the oxidation kinetics and the microstructure of the oxide scales and interconnect materials. A clear correlation between increasing temperatures and increasing oxide growth rate constants kp,w can be demonstrated in all materials. This interrelation results in thicker surface oxide scales. A comparison of surface oxide thicknesses after 1000 h oxidation in reducing atmosphere show an increase between 725°C and 875°C with a maximum factor of about 3.8 in ferritic steels and 2.4 for CFY. The comparison of kp,w values at temperatures between 725 and 875°C show acceleration factors of 8,2-11.9 in ferritic steels and 2.3 for CFY. No significant dependence of the oxide growth rate in ferritic steels and CFY by variation of the water content in fuel gas for concentrations within H2/H2O=87/13 vol.% and H2/H2O= 50/50 vol.% was found. The structures of the oxide layers are specific for each material and consist mainly of Cr2O3 (CFY) and Cr2O3/(Cr,Mn)3O4 in ferritic steels with different element distributions and thickness ratios. Beside this, the zone of an inner oxidation of MICs with the Al-, Si- and Ti- rich oxide inclusions can be seen in ferritic samples, whose microstructures differ depending on analyzed materials and temperatures. In Nb containing materials (Crofer H, AISI 441), Silicon precipitations as Laves phases as well as continuous SiO2 scales were observed beneath the formed Cr2O3 top layer. Furthermore, an increase of the internal oxidation zone was identified with increasing temperatures in ferritic steels. In CFY, the internal oxidation zone with Cr2O3 inclusions in the material bulk remains almost constant over 1000 h oxidation in the tested temperature range. Electrical measurements in the dual gas atmosphere reveal also an increase of resistance within 1000 h material exposure under anode gas conditions. Comparison of these results suggests that materials, forming such isolating SiO2 layers, show a higher resistance increase. The results of the oxide scale formation under the anode gas conditions in tested Crofer 22 APU and CFY samples are compared with data derived from the real SOFC stacks. Only minor differences in the oxide thicknesses were observed between the gas inlet and gas outlet (region with higher water content) at the anode side of the stacks. This behavior is similar to the results of our artificially oxidized MIC samples in which we also have not found a clear evidence for an increase in oxide scale growth rate with higher water content in the humidification range tested. Such effects can indicate, that the reactions at phase boundary are dominated by solid-state diffusion and the rate supply of adsorbed oxygen species (from H2O) is not the limiting factor (in the tested H2O concentrations range).
Read full abstract