The cyclic stress-strain response of copper at low strains—i. Constant amplitude testing

Abstract

A controversy has arisen over whether or not the cyclic stress-strain (CSS) curve for polycrystalline copper can be correlated with that for monocrystals, particularly in showing a plateau and a fatigue limit at an equivalent stress and strain (equivalence being expressed via Taylor's factor). Bhat and Laird suggested this correlation but Mughrabi opposed it on the grounds that interpretetion of the mechanical data was equivocal and he did not believe persistent slip band (PSB's) would exist in the bulk of polycrystals. To help settle this controversy, low strain tests under both constant strain-control and constant stress-control (Part I) have been carried out in polycrystals as well as extensive observations of the dislocation structures by TEM. It is concluded that the strain-control curve does not show a true plateau, in agreement with the views of Mughrabi. However, a definite change of slope occurs at a stress and strain corresponding to the fatigue limit and initial formation of PSB's in monocrystals, in agreement with the views of Bhat and Laird. Furthermore, the stress-control CSS curve tends strongly to a plateau, differing from the strain-control CSS curve at low strains; also dipolar wall structures analogous to those of monocrystalline PSB's but not so regularly ladder-like, are found in the bulk of polycrystalline copper. The dislocation structures corresponding to the different CSS curves in load and strain control are documented, and reasons for history dependence in CSS response are provided. Thus a close correlation between the CSS curves and cyclic response of polycrystals and monocrystals is demonstrated. In a companion paper (Part II), similar correlation is demonstrated for tests conducted under variable amplitudes.