Observation of twin boundary migration in copper during deformation

Abstract

A previous investigation produced evidence that twin boundaries in annealed copper were a significant source of dislocations during the initial stages of plastic flow. The character of the dislocation source was unknown, but it was hypothesized that twin boundaries could be non-regenerative dislocation sources that would cause migration of the boundary during plastic deformation. Channel die deformation and intermittent orientation imaging were performed on split specimens of pure copper in an attempt to observe twin boundary migration. Approximately 15% of the twin boundaries were observed to migrate beyond that expected from the imposed strain. The data support the hypothesis that twin boundaries can serve as dislocation sources.

Introduction

In an earlier study, the authors reported observations on the influence of grain size on the flow stress behavior of polycrystalline copper [1]. The results of the study are summarized as follows: (1) the grain size including annealing twins only affected the onset of plastic flow showing no influence on strain hardening; (2) the flow stress dependence on grain size was associated with a reciprocal boundary (twin and grain) and not on an inverse square root dependence; (3) TEM images after small amounts of plastic deformation showed dislocation debris including loops along twin boundaries with few dislocations evident on random boundaries; and (4) orientation imaging showed unmistakably that the overall fraction of Σ3 boundaries were reduced during deformation. The strengthening from grain size, in particular the area of annealing twin boundary per unit volume, is associated with the activation of dislocation sources at the initial stages of plastic deformation. Since the grain size had no influence on the subsequent strain hardening of copper, it is evident that dislocation–dislocation interactions control hardening behavior. Dislocation interactions with grain and twin boundaries do not contribute significantly to strain hardening at larger strains where a refined dislocation network controls mechanical response of the metal.

The question that has not been significantly resolved is the character of the dislocation sources during the initial stages of plastic deformation. Our prior evidence [1] supports the interpretation that the annealing twin boundaries can be significant sources for dislocations. Other investigators have reached similar conclusions [2], [3], [4], [5].

The experiments described in this present study on polycrystalline copper are designed to observe twin boundaries at incremental stages of deformation with the intent to determine whether the imposed strain causes the boundaries to migrate or somehow “unravel.” Observation of the movement of an apparent coherent twin boundary would confirm that they are serving as non-regenerative dislocation sources (as opposed to the more common Frank–Read source that is regenerative in nature). Similar to observations near any dislocation source, we also expect that lattice curvature behind the migrating twin boundary will be significant since dislocations are emitted into the lattice. Measurement and quantification of these phenomena are the objective of this paper.