Abstract
A theory of steady-state creep at high temperature and stress is based on the climb of dislocations limited by the diffusion of vacancies from them. The creep rate is found to be approximately (MnL3)D0 exp[− (W+E+U−ΔW)/kT], where M is the number of dislocation sources per unit volume, n is the number of dislocations of length L per unit area, and D0 exp (−W/kT) is the coefficient of self-diffusion. The number of jogs in the dislocations is determined by U, and E is an activation energy for formation of vacancies, due to relaxation of the lattice. The reduction in vacancy formation energy near a dislocation ΔW is estimated from the interaction energy of the climbing dislocations. A comparison of experimental data for Zn and AgBr suggests that E is a significant contribution to the activation energy for ionic crystals but not for metals. The predicted increase in vacancy concentration during creep is too small to give observable effects. It is concluded that any observed enhancement of diffusivity during creep is due to increased mobility of defects.