On the structure of grain boundaries in metals

Abstract

This review consists of an introduction (Section 1) and two main sections (Sections 2 and 3). In Section 2 the following atomic models proposed for the equilibrium structure of grain boundaries in metals are dealt with. 1. (1) First, dislocation (disclination) models describing the structure of large-angle grain boundaries in terms of a dense array of dislocations or disclinations are considered. Because of the small dislocation spacing in large-angle boundaries, particular attention has been paid to the core properties of these dislocations. The various methods employed to calculate core properties in dense dislocation arrays are critically discussed. 2. (2) In the development of coincidence models the emphasis has been shifted in recent years from models based on the existence of lattice or boundary coincidence atoms to models describing the structure of grain boundaries in terms of a relaxed two-dimensional periodic arrangement of atoms, none of which, in general, occupies coincidence sites. This development and the resulting models are critically reviewed. Particular attention is focused on the present physical understanding of the existence of low energy boundaries and the crystallography of interphases (the O lattice concept). 3. (3) Plane-matching models are discussed in view of the existing experimental evidence. 4. (4) In terms of polyhedral unit models the structure of a grain boundary is interpreted as a two-dimensional array of one or several types of specific atomic configurations. The historical development of this idea is summarized and the various types of models proposed so far are discussed with regard to the advantages and limitations of this description. Section 3 is concerned with the structure of non-equilibrium defects in grain boundaries. Two defects are considered: vacancies and dislocations. Recently published observations suggest that the structure of grain boundary vacancies may be different from lattice vacancies. The free volume associated with the vacancy core may be smeared out (delocalized) in the plane of the grain boundary. The tendency for delocalization seems to be lowest in low energy boundaries. The three models proposed for the structure of dislocations in grain boundaries are reviewed. In the dissociation model, any extrinsic grain boundary dislocation is assumed to dissociate into the corresponding DSC boundary dislocation. The delocalization model proposes that the core width of grain boundary dislocations depends on the boundary structure in the sense that the core width increases and the long-range strain field weakens with increasing boundary energy. Because of the resulting core overlap of neighbouring boundary dislocations, the delocalization may limit the physical significance of boundary dislocation models to those boundaries with localized boundary dislocations and/or widely spaced boundary dislocations. On the basis of experimental observations, the strain-sharing concept describes the structure of an extrinsic dislocation in terms of structural rearrangements of the neighbouring (structural) boundary dislocations so that the strain field of the extrinsic defect will be effectively spread out.