Surface relief and the mechanism of a phase transformation

Abstract

Surface relief of the invariant plane strain type is invariably observed during martensitic transformations. However, some diffusion-controlled transformations also exhibit this effect. The observation of a surface relief effect is therefore no proof for a martensitic transformation mechanism. The same reasoning applies to the formation of groups of variants that accomodate each other's strain. Self-accommodating groups of four variants of martensite plates have been observed in several martensites but they are also found in precipitation reactions. While these observations cannot be used as proof for a martensitic mechanism, neither does the presence of structural ledges in a reaction interface necessitate diffusion. Both these arguments have been used by different sides in a recent controversy (1,2), and the present paper is an attempt to clarify the situation. Starting with the simple example of AlAg, the mechanism of surface relief in a diffusional fcc-hcp transformation and its relation to ledge growth will be examined. Applying the same mechanism to increasingly complex cases the relation to structural ledges, fcc-bcc transformations and finally a general martensite interface will be developed.