I am involved in projects pertaining to the modeling and simulation of materials deformation, shock dynamics, and radiation damage. The main focus of the work is at the atomistic level, where I have a long interest in developing models that go into the simulations. The techniques that I employ are those of equilibrium and nonequilibrium , molecular dynamics, and Monte Carlo sampling techniques. I employ several advanced computational architectures including Blue Gene and cell-blade platforms. Atomistic model development requires extracting critical elements of electronic structure theory for implementation at this coarser level.
One example of materials deformation is 3D study of slip transmission across an aluminum grain boundary generated by a dislocation pile-up. Two grains of aluminum are stacked on top of one another in prescribed orientations. Two dislocations that are being driven into the boundary by an external load appear in the lower left quadrant of the Figure. One other dislocation in the pile-up has already reflected off of the boundary (lower right quadrant). Two more dislocations have transmitted, one on each of the available slip planes in the upper grain. Dislocation transmission leads to deformation of a macroscopic body.
These simulations were performed on the Cerrillos computer platform at Los Alamos National Laboratory, with the SPaSM molecular dynamics code. Collaborators include Tim Germann, Jian Wang, and Dick Hoagland. The work was supported by the LDRD Program Office.
For years, many research groups have been trying to develop atomistic models for ceramics based on the notion that the energies of the ions making up the ceramic can be represented by a quadratic polynomial in their net charges (quadratic term in red, marked "IM" in the Figure below). Results from density functional theory indicated that this was not always the case but did not say what to use instead of the quadratic polynomial.
A new theoretical analysis gives a comprehensive answer: The linear term of these models is correct, but the quadratic term is not (black curve for bonded and metallic interactions, and blue curve for debonded and insulating.) But the analysis gives another insight. One special property coming out of the density functional theory analysis is responsible for regulating charge flow. The original models do not have this property. This is because there are energy contributions that do not appear in the original models with much different charge dependencies. A charge dependence, derived from the analysis, shows the variation in charge-flow regulation between metallic and insulating (Figure below) conditions.
The intersection of the colored lines with the straight, black line shows the charge state where flow stops. The purple curve represents extreme attenuation of charge flow, while the turquoise curve represents facile flow.
This work was supported by the LRDR and CMIME Programs.
- Richard Hoagland
- Anand Kanjarla
- Ricardo Lebensohn
- Tongsik Lee
- Ben Liu
- Stephen Niezgoda
- Steve Valone
- Jian Wang