Date of Award

2018

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Materials Science & Engineering

First Advisor

David J. Srolovitz

Abstract

Many materials of scientific and industrial interest are polycrystalline in nature. That is, they are composed of microscopic crystalline grains, delineated by grain boundaries (GBs). The properties and utility of these materials are highly sensitive to the microstructure, composed of GBs and other defects. This presents both an opportunity and a challenge; one may tune the properties of a material by altering the microstructure, but these properties will change if the microstructure subsequently evolves. Typically, it does. Grain growth (the increase in mean grain size over time) is a cornerstone problem in Materials Science; the classic solution, the analogy of a coarsening soap film, has stood for decades. However, it is now understood that GBs migrate via the nucleation and migration of disconnections, step defects with dislocation character. Disconnections induce shear as GBs migrate, leading to stress generation and potentially grain growth stagnation. Similarly, GB triple junctions (TJ) migrate by the reaction of disconnections exchanged between them and their constituent GBs. This is only possible with special combinations of disconnections, such that the GBs continuously meet along the TJ line and the total Burgers vector cancels. How have classical grain growth theories been successful despite neglecting these considerations?

In this thesis, the disconnection theory will be placed in context and rationalized with MD grain growth observations. It will be shown that disconnection-mediated GB migration leads to stagnation, but that this can be overcome if GBs can nucleate multiple disconnection varieties. A kinetic model of multi-mode GB migration will be presented; it captures the full range of migration behavior observed in MD. The analysis will extend to TJ migration, which is also facilitated by multi-mode GB migration. When multiple modes are unavailable, TJs may emit crystal defects (e.g., dislocations and twins) to assist migration. This results in a rich and complex microstructure; after forming, twins may interact and form new composite twin junction structures. These will also be examined. Finally, one of the gaps in disconnection theory is a knowledge of energy barriers for disconnection nucleation and migration. These are difficult to measure and are only known for certain disconnection modes of selected GBs. To address this, a machine learning method probing atomic-level rearrangements will be presented. This method characterizes an atomic-level structural quantity, Softness, measuring the likelihood of local atomic rearrangement. This technique was originally developed for the analysis of disordered materials, but is here applied to polycrystals. This may be a useful tool for studying the atomic dynamics of GBs and determining representative energy barriers for the refinement of disconnection models.

This work, in sum, aims to extend the science of grain boundary kinetics from idealized, domesticated grain boundaries to a full theory of grain boundaries in the wild.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS