Investigation of dislocation behavior in micron and sub-micron thin films

Abigail Hunter, Purdue University

Abstract

Plastic deformation in crystalline materials is mediated by dislocation motion and their interaction with defects, such as second phase particles, dislocations, grain boundaries and voids. In addition, grain boundaries, free and passivated surfaces have a significant impact on the evolution of dislocations and their intricate structures. In polycrystalline materials, the influence of dislocation motion and interactions results in unique mechanical properties, such as high yield stress and fracture strength and a dependency on grain size. It is observed that for an average grain size in the micron and sub-micron regime, the yield stress increases as the grain size decreases following a power law. This size effect is known as Hall Petch effect. A reliable computational model that describes the mechanical response and failure mechanisms of micron and sub-micron scale devices should incorporate these size effects. A three-dimensional phase field dislocation dynamics model (3D PFDD) is developed. This is a dislocation based plasticity model that accounts for the motion and interactions of individual dislocations with material defects and interfaces, such as obstacles, and grain boundaries. This model is a valuable and efficient research tool that will help to understand plastic deformation on the mesoscopic level, bridging the gap between microscopic and macroscopic studies. For the research presented here, this model is used specifically to understand and simulate dislocation behavior in fcc (face-centered cubic) metal thin films, similar to those used in micro-electro-mechanical systems (MEMS). Incorporating microstructure, such as grain boundaries, is key to accurately predicting deformation behavior in any system. Plastic deformation is affected by both the thickness of the film layers and by the resolution of the film's internal microstructure. In MEMS devices and components that are generally on the micron scale (hundreds of microns in size), the internal microstructure, such as grain size, shape, etc., is on the nano-scale (ranging from tens to hundreds of nanometers) (Chen et al., 2007; Van Swygenhoven et al., 1999a; Van Swygenhoven et al., 1999b). At this length scale, implementing partial dislocations into the PFDD code is important for modeling the correct dislocation behavior and hence in accurately predicting performance, reliability, and lifetime. The investigation of size effects is a primary goal of this research, and has been achieved through the representation of the microstructure and partial dislocations in to the 3D PFDD model. This requires several steps detailed in the following thesis, including: validation against experiments, and verification of the model with analytical solutions and other simulation tools; parallelization of the code to allow for large-scale simulations on high performance computing resources; and collaboration and connections to simulation tools at length scales both above and below the mesoscale. Simulations have been completed and analyzed for Nickel thin films undergoing plastic deformation. Overall performance and efficiency of the parallel algorithm is discussed, along with comparison to analytical solutions and modified continuum models. In addition, the impact of grain size on yielding, and hardening behavior is analyzed for several active slip systems. Finally, the implementation of partial dislocations is presented with simulation results. Understanding dislocation dynamics is particularly important in the study of plastic deformation in small-scale crystalline structures. At micron and sub-micron scales, the study of the impact of dislocations on material properties, reliability, and failure becomes increasingly important as more and more applications emerge. At this scale, phenomena such as grain sliding, grain boundary diffusion and migration, and the interaction of dislocations with grain boundaries, obstacles, and surfaces play a prominent role in the evolution of plastic deformation. These deformation mechanisms are not accounted for in classical models that describe bulk materials and, hence, accurate deformation behavior of small components is hard to predict (Hunter and Koslowski, 2008). This 3D PFDD model seeks to fill this gap with an accurate and computationally efficient simulation tool for tracking and predicting dislocation behavior. (Abstract shortened by UMI.)

Degree

Ph.D.

Advisors

Koslowksi, Purdue University.

Subject Area

Computer Engineering|Mechanical engineering|Materials science

COinS