AbstractMany high–temperature creep tests are performed on low–chromium, ferritic steels in an uncontrolled atmosphere. Examination of creep rupture specimens of 0·5Cr–0·5Mo–0·25V steel tested in air has shown that decarburization accompanies oxidation and is an important factor in accelerating the failure of creep tests in air. Similarly, pre-aging in air reduces the creep life more than pre-aging in a capsule. There is also evidence that decarburization is accelerated during straining. Measurements of surface carbon contents in 10 mm thick blocks heat treated in air at 600–700°C have given an apparent activation energy for decarburization of about 250 kJ mol−1, at least twice that for carbon diffusing in ferrite. However, this value is still below that for creep, so the influence of decarburization on creep life is expected to increase at lower temperatures. Structural observations are discussed in relation to loss of carbon and are related to creep behaviour. Secondary precipitation was observed after low-temperature treatments in aged encapsulated specimens, but not in specimens aged in air. This is attributed to the loss of carbon in the air aged specimens, which also showed a decrease in the M3C content. The iron content of M3C particles depends on carbon content as well as aging time.MST/40