The increased use of magnesium alloys in the automotive industry continues to drive intensive efforts to optimize/tailor their properties to various potential applications in cars. Finding an optimal balance between material ductility and strength is one critical area where new fundamental and technological insights are needed. Ductility of hexagonal close-packed (HCP) materials is largely defined by multiplication and motion of <c+a> dislocations that are only marginally stable and prone to spontaneous dissociation into (less valuable) <c> and <a> component dislocations. Kinetic stability of <c+a> dislocations under mechanical loads potentially can be improved through alloying. To explore the potential of alloying to stabilize <c+a> dislocations in magnesium, we employ two unique computational capabilities developed recently at Lawrence Livermore National Laboratory. Our plan involves using electronic structure calculations to predict—with chemical accuracy—how much the balance of core energies in the <c+a> = <c> + <a> dislocation reaction can be shifted through alloying and employing dislocation dynamics simulations to accurately model the effect of changes in the predicted core energy balance on strength and ductility of magnesium single crystals strained along the 0001 axis. We are exploring several potential alloying elements to predict and quantify their effects on the 0001 ductility of magnesium. The goal is to gain quantitative and semi-quantitative insights into the physics of dislocation core in magnesium alloys and develop practical recommendations for alloy design and optimization for automotive applications.
Exploring alloying for modifying dislocation core energetics in magnesium alloys demands HPC capabilities that, while cutting-edge now, should become more widely available in the near future. Presently, only sparse computational sampling of alloy compositions is practical, i.e., on order of 20–50 alloy compositions over two to three years.
Accurate predictions of dislocation core structure and energy require large-scale electronic structure calculations to discriminate useful alloying trends to within 1 meV accuracy (per computation supercell). This capability is recent and unique to LLNL. The laboratory also pioneered dislocation dynamics simulations of HCP metals that can be used to quantify the effects of alloying on crystal ductility and strength.
Density functional theory calculations constitute a large portion of computational effort directed toward computational materials engineering in high-performance computing centers across academia, government laboratories, and industrial research and development centers. As such, adoption of LLNL’s new capability for accurate predictions of the effects of alloy chemistry on dislocation properties likely will be rapid and widespread. In addition, LLNL develops and maintains the dislocation dynamics code, ParaDiS, now available in the public domain. New dislocation dynamics simulation capabilities for HCP metals will be added to the public version of ParaDiS within the next two years.
Name: Vasily Bulatov
- Wei Cai, V. V. Bulatov, J.P. Chang, Ju Li and Sid Yip. “Dislocation core structure and mobility”, (2003) in: Dislocations in Solids, volume 12, edited by F.R.N. Nabarro, pp. 1-80.
- Wei Cai, V. V. Bulatov, J.P. Chang, Ju Li and Sid Yip, “Anisotropic Elastic Interactions of a Periodic Dislocation Array” (2001), Phys. Rev. Lett. 86, 5727 (2001).
- S. Aubry, M. Rhee, G. Hommes, V. V. Bulatov and A. Arsenlis, " Dislocation dynamics in hexagonal close-packed crystals" (submitted to J. Mech. Phys. Solids, 2015).