Energy barriers controlling creep processes in metals with special reference to β-cobalt

Abstract

Abstract The minimum, second stage, creep rate ϵ of polycrystalline cobalt (99·999%), subjected to constant stresses in vacuo at 430–750°C, is shown to obey the relation ϵ = A(T) exp [(-H + qσ)/kT] over the entire range of strain rates employed, extending from 10-9 to 10-4 per sec. As in copper and α-brasses, two sets of the parameters A, H and q are however required to describe any given log ϵσ isotherm, depending on whether the tensile stress σ is greater or less than a critical value σc(T) at which an abrupt change occurs in the creep mechanism. The increase of the activation volume 1/2q with temperature is governed by an activation energy Q = 0·55 ev which, as in other metals, is close to the theoretical value of the interaction energy between the hydrostatic stress field of dislocations and vacancies. H, determined by thermal cycling, was found to be equal to 2·2 ± 0·25 ev and 1·9 ± 0·25 ev for σ < σc and σ > σ;c respectively, but at about 550°C rose steeply with temperature in both cases to a common constant level of 2·9 ev, numerically equal to H sd, the activation energy of self-diffusion. The first two values are ascribed to the movement of jogs in mixed dislocations, the respective barriers being H sd - Q and H sd - 2Q. The barrier at higher temperatures is considered to arise from the non-conservative migration of jogs in dislocations of predominantly screw character. Corresponding energy spectra of other metals are reviewed.