The high temperature oxidation of metals and alloys is slowed to a useful extent by the presence of certain elements or their oxides in the metal or on its surface. Particles of the reactive element oxide are found in part of the oxide scale and, by an unknown mechanism, retard the diffusion of metal through oxide grain boundaries. Changes in texture are also observed, which might affect geometry of individual grain boundaries. In this study the rotational geometry of grain boundaries in nickel oxide scales has been characterized both in the presence of particles of cerium dioxide and in their absence. Proximity of the relationship of neighbouring grains to a coincidence orientation could have an important influence on diffusion rate. Ways of testing this are compared and a new method for analysing orientation data from electron diffraction is presented, which allows statistically sound conclusions to be drawn. The frequency of occurrence of coincidence orientations in the presence of CeO, is little greater than random, so one possible mechanism for the 'reactive element effect' may be rejected.
I N T R O D U C T I O NThe oxidation of metals and alloys at high temperature has important economic consequences. To prolong the life of artefacts such as furnaces, jet engines and fuel elements for gascooled nuclear reactors, relatively high levels (>20%) of expensive chromium have to be used to give a protective Cr203 scale (Wallwork, 1976) and even then, loss of the scale by spalling can limit operating temperatures to well below the limits imposed by mechanical strength alone. One palliative measure is the introduction of a small proportion of certain reactive elements or their oxides into the metal by alloying or dispersion, or on its surface as a coating. This improves performance by reducing both spalling and the oxidation rate itself (Whittle 8t Stringer, 1980). The lanthanides, titanium, hafnium and thorium are all effective to varying extents. Similar results can be obtained by ion implantation. A common factor appears to be the low solubility of the reactive element in the native oxide (Caillet et al., 1983) and Ecer et al. (1982) have suggested that its segregation to grain boundaries blocks cation diffusion. There is a good deal of evidence that grain boundary transport of cations dominates oxidation processes of a variety of metals and alloys at temperature of practical importance (Stringer, 1973;Hughes et al., 1984). 0 1985 The Royal Microscopical Society 22 1