It was recalled that the observed features of yield drops during work-softening supported the suggestion that the structure in a slip band was heterogeneous in metals such as Al and Cu. The centre of the band was almost empty of dislocations, while the latter were concentrated round its boundary; particularly as walls of edge dipoles at the ends. It was proposed that work-hardening was due to the blocking of these dislocations by a forest of obstacles having small activation volumes and energies. Thermal energy permitted them to cut through the obstacles, and penetrate further into the walls; thus reducing the back-stress and permitting Frank-Read sources within the bands to become active at lower applied stresses. The yield drop was explained in terms of an overshoot, of the stress required to drive the sources, which resulted from the thermally activated cutting.

A Theory of Work Softening. A.H.Cottrell: Philosophical Magazine Letters, 2001, 81[1], 23-8