Effect of SO sub(2) on oxidation of metallic materials in CO sub(2)/H sub(2)O-rich gases relevant to oxyfuel environments

In an oxyfuel plant, heat exchanging metallic components will be exposed to a flue gas that contains substantially higher contents of CO sub(2), water vapor, and SO sub(2) than conventional flue gases. In the present study, the oxidation behavior of the martensitic steel P92 was studied in CO sub(2)- and/or H sub(2)O-rich gas mixtures with and without addition of SO sub(2). For this purpose, the corrosion of P92 at 550 degree C up to 1000h in Ar-H sub(2)O-SO sub(2), Ar-CO sub(2)-SO sub(2), Ar-CO sub(2)-O sub(2)-SO sub(2) and simulated oxyfuel gas (Ar-CO sub(2)-H sub(2)O-O sub(2)-SO sub(2)) was compared with the behavior in selected SO sub(2)-free gases. The oxidation kinetics were estimated by a number of methods such as optical microscopy, scanning electron microscopy with energy and wave length dispersive X-ray analysis, glow discharge optical emission spectroscopy, X-ray diffraction as well as transmission electron microscopy. The experimental results revealed that the effect of SO sub(2) addition on the materials behavior substantially differed, depending on the prevailing base gas atmosphere. The various types of corrosion attack affected by SO sub(2) could not be explained by solely comparing equilibrium activities of the gas atmospheres with thermodynamic stabilities of possible corrosion products. The results were found to be strongly affected by relative rates of reactions of the various gas species occurring within the frequently porous corrosion scales as well as at the scale/gas- and scale/alloy interfaces. Whereas SO sub(2) addition to Ar-CO sub(2) resulted in formation of an external mixed oxide/sulfide layer, the presence of SO sub(2) in oxyfuel gas and in Ar-H sub(2)O-SO sub(2) resulted in Fe-sulfide formation near the interface between inner and outer oxide layer as well as Cr-sulfide formation in the alloy. In the latter gases, the presence of SO sub(2) seemed to have no dramatic effect on oxide scale growth rates.