Study of size effects in thin films by means of a crystal plasticity theory based on DiFT

TL;DR

This paper develops a continuum crystal plasticity theory based on DiFT to analyze size effects in thin films, capturing dislocation interactions and boundary layer phenomena observed in discrete simulations.

Contribution

The study introduces a novel continuum dislocation field theory (DiFT) for size effects in thin films, incorporating pair correlations and boundary layer behavior.

Findings

Size effects arise from dislocation boundary layers near interfaces.

The continuum theory predicts stress responses consistent with discrete simulations.

Boundary layers depend on slip orientation and loading, not film thickness.

Abstract

In a recent publication, we derived the mesoscale continuum theory of plasticity for multiple-slip systems of parallel edge dislocations, motivated by the statistical-based nonlocal continuum crystal plasticity theory for single-glide due to Yefimov et al. (2004b). In this dislocation field theory (DiFT) the transport equations for both the total dislocation densities and geometrically necessary dislocation densities on each slip system were obtained from the Peach-Koehler interactions through both single and pair dislocation correlations. The effect of pair correlation interactions manifested itself in the form of a back stress in addition to the external shear and the self-consistent internal stress. We here present the study of size effects in single crystalline thin films with symmetric double slip using the novel continuum theory. Two boundary value problems are analyzed: (1) stress relaxation in thin films on substrates subject to thermal loading, and (2) simple shear in constrained films. In these problems, earlier discrete dislocation simulations had shown that size effects are born out of layers of dislocations developing near constrained interfaces. These boundary layers depend on slip orientations and applied loading but are insensitive to the film thickness. We investigate stress response to changes in controlled parameters in both problems. Comparisons with previous discrete dislocation simulations are discussed.