“Our work presents a scientific framework for investigating the material behavior at those relevant scales, providing means for predicting mechanical properties, material performance and failure mechanisms.”

Over the past few decades, there has been a tremendous amount of effort devoted in the materials engineering science and mechanics communities towards developing a better understanding of the way crystalline materials behave under various conditions and over wide ranges of length and time scales. We now understand that deformation and strength of metals and semiconductors are determined to a large extent by underlying mechanisms involving various crystal defects, such as vacancies, interstitials and impurity atoms, dislocations, grain boundaries, heterogeneous interfaces and microcracks, chemically heterogeneous precipitates, twins, and other strain-inducing phase transformations. Most often, however, "dislocations" define the materials behavior, either as the dominant carriers of permanent deformation or through their interactions with the other strain-producing defects. Therefore, accurate prediction of the strength of materials requires modeling the dynamics of dislocations and the manner in which they interact with other defects. However, these defects span a wide range of length and time scales. Our work presents a novel framework for investigating the material behavior at the nano-to-micro scales and involves the prediction of mechanical properties of metals through the use of so-called "three-dimensional dislocation dynamics simulations."

We have developed a new framework in computational materials science involving the establishment of new techniques and methodologies that are quite different and more complex than those involved in the classical finite element analysis and molecular dynamics typically used by scientists and engineers. It involves a numerical model for discrete dislocation dynamics code called micro3D coupled with a continuum model which provides scientists and engineers with the ability to address the complexities of deformation. The model micro3D has been used to address and examine a broad range of problematic phenomena in materials science and mechanics. The work has appeared prominently in many international journals, most notably in Nature. It is being used by many engineers and scientists at leading national laboratories and universities.

In recent years, the miniaturization in electronic components, micromechanical devices and sensors has reached the point where the volumes of materials present are so small that the properties of the material are significantly different from those measured in the bulk. Apart from the size effect, the novel processing routes for nano- and micro- components produce different nano-structures and thus, the material properties are different than those of the bulk material. Our work presents a scientific framework for investigating the material behavior at those relevant scales, providing the means for predicting mechanical properties, material performance and failure mechanisms.

This is part of a larger effort being carried out at Lawrence Livermore National Laboratory and at a number of universities, including Washington State University, to develop physically-based multiscale models that can predict the behavior of matter under an extremely wide range of conditions and at length scales ranging from nanometers to macrometers. This effort aims at building a hierarchy of multiscale models, integrating ab initio electronic structure calculations, molecular dynamics simulations, kinetic Monte Carlo, microscopic simulations, mesoscopic analyses, and macroscopic calculations over the relevant length and time scales. Our work aims at linking the molecular dynamics scale with the mesoscopic scale, predicting the evolution of the relevant microstructure and linking it to a quantitative prediction of mechanical properties through the use of three-dimensional dislocation dynamics simulations. An important component in this effort is the integration within experiments, both to validate the modeling and to assist in their interpretation.