5542 - MULTI-SCALE MODELING OF DEFORMATION IN POLYCRYSTALLINE THIN METAL FILMS

alexander hartmaier,m buehler,huajian gao
2013-01-01
Abstract:The time-dependent irreversible deformation of thin metal films on substrates is investigated on various levels of modeling. The work focuses on polycrystalline films without cap layer, where diffusive processes play an important role. A continuum level model of constrained grain boundary diffusion describes how surface and grain boundary diffusion lead to the relaxation of tractions along the grain boundaries. If the film is perfectly bonded to the substrate, the relaxation of tractions along the grain boundary leads to a crack-like stress concentration close to the film/substrate interface. By virtue of this stress concentration, dislocations on glide planes parallel to the thin film are subject to a driving force that does not exist in the case of purely bi-axial loading. This model gives a theoretical basis for experimental observations of such parallel glide processes. To obtain a better understanding of the microscopic phenomena associated with the this creep process, we conducted large scale molecular dynamics (MD) simulations, as well as discrete dislocation dynamics (DDD) studies on intermediate length scales. The results of these numerical simulations reveal that nucleation of glissile dislocations at the tip of the diffusion wedge can be described with a critical stress intensity factor criterion. The validation of this criterion for dislocation nucleation by MD simulations allows transferring this information onto the DDD length and time scales. The deformation behavior of thin films is investigated with DDD simulations on length scales from nanometers to microns. If only dislocation glide is active, DDD simulations with a constant density of dislocation sources show that the flow stress of thin films scales linearly with the inverse film thickness. Furthermore, the DDD simulations reveal that, upon activation of the creep mechanisms in ultra-thin films the size scaling of the flow stress breaks down. Instead, such films show a thickness-independent flow stress. These findings agree well with experimental observations.
What problem does this paper attempt to address?