Computational materials science uses computer models of atoms and atomic interactions to predict material properties. The main focus of my group is on structural materials, which directly impacts energy efficiency by improving processing capabilities of lighter materials (leading to weight reduction in automotive and aerospace) or improving high temperature properties for materials in energy production (leading to increased operational temperatures for turbines in aerospace and reactors).
My group studies chemical effects on mechanical properties of materials, such as plasticity and phase transformations. I've calculated phase transformation mechanisms in Ti, and dislocation/solute interactions in Mo and Al. In addition to those materials, I currently study Fe, Al, and Ni-based superalloys. I use atomistic methods (electronic structure, tight-binding, classical potentials) coupled to larger length-scale models (continuum elasticity, statistical mechanics). This work is both connected to applications, and involves improving and developing new computational techniques.
I am looking to hire motivated, interested Ph.D students. Previous computational science experience is not necessary. If you are a UIUC student, contact me if you're interested in joining my group.
Simulations model atomic interactions. At the highest end of accuracy are density-functional methods, which treat electronic bonding between atoms; with these ab initio methods, atomic chemistry can be changed at will. Those calculations are often limited to 100-500 atoms; my group studies approaches to extend that applicability.
Lattice Green function method. Typical density-functional calculations require periodic boundary conditions, which often require very large computational cells for defect structures. This inefficiency can be resolved by using flexible boundary conditions that accurately feedback the correct long-range strain field of a defect. This method makes simulations of single, isolated dislocations with density-functional theory.
Atomistic potentials. An alternate approach is to build a simpler model of interatomic interactions; the model has adjustable parameters that can be optimized to a database of accurate property calculations. Optimizing non-linear models is non-trivial; I'm interested in new approaches that help to automate this process by extracting more model information.
Publications:
Using direct density-functional theory calculations of solutes interacting with dislocations allows accurate models of dislocation motion in real alloys. In the case of Mo, some 5d solutes can make Mo softer, but only for a range of temperatures and compositions. An accurate quantitative model of solid-solution softening was lacking for the nearly 50-years since it's discovery in bcc metals. A model for thermally-activated dislocation motion in the presence of solutes accurately captures the flow stress and hardness with temperature, solute chemistry, and composition. The direct calculations also showed that the chemical bonding in the center of a dislocation is different than the bulk environment, and this difference is important.
The approach for examining dislocation/solute interactions is being applied to the problem of Mg solutes pinning dislocations in Al. The pinning by a "Cottrell atmosphere" of solutes is governed by the attraction of Mg solutes to the tensile strain field of an edge dislocation. Accurate density-functional calculations provide interaction energies all the way to the core of the dislocation. The accurate modeling of Al dislocation/Mg solute interaction is important for dynamic-strain aging: mechanical behavior that produces shear-bands, and limits the formability of Al alloys.
Publications:
Martensitic phase transformations are structural phase transitions that occur without diffusion; they can propagate near the speed of sound. The formation of martensitic phases are key to geophysical processes, strengthening in steels, and the shape-memory effect. In Ti, the high pressure omega phase embrittles alpha (hcp) Ti alloys. However, as little as 1at.% oxygen can push the transformation pressure from 9GPa up to more than 35GPa. Modeling changes in the transformation with solutes and interstitials requires (1) the atomistic mechanism, and (2) calculation of changes in the energy barrier. A new algorithm was developed that permits systematic generation and sorting (by energy barrier) of transformation pathways; this algorithm found a new low energy barrier pathway for Ti alpha-omega called TAO-1. The change in energy barrier from different solutes predicts changes in transition pressure for engineering Ti alloys.
Publications: