Abstract
Grain boundary sliding is necessary during diffusion creep to maintain continuity across the grain boundaries. In this paper it is shown, by use of a two-dimensional hexagonal array, that grain boundary sliding will contribute additional axial strain to the specimen as predicted by R. N. Stevens (1971). It is shown, however, that there is a difference between the macroscopic effect of this type of sliding and that of ‘ordinary’ grain boundary sliding, in that the former, named Lifshitz grain boundary sliding, does not increase the number of grains along the length of the specimen, whereas, the latter, named Rachinger grain boundary sliding, does. Considering only diffusion creep, it is concluded that where egb(L) is the strain due to Lifshitz grain boundary sliding, ed is the strain a polycrystalline aggregate would experience if only stress-directed diffusion were to occur without the necessary grain boundary sliding in which case grain boundary separation would occur, egb is the grain strain et is the total strain and egb(R) is the strain due to Rachinger grain boundary sliding. The prediction of R. N. Stevens as to the fraction of strain resulting from grain boundary sliding may be more precisely stated in the form T gb(L) = e gb(L)/e, It is this fraction which is predictable only from considerations of grain geometry.

This publication has 6 references indexed in Scilit: