Following is a brief summary of the invited presentations from the joint APS/TMS symposium "Frontiers in Computational Materials: Session I" held in New Orleans at TMS Annual Meeting 2008.
• M. Scheffler, Fritz-Haber-Institut der Max-Planck-Gesellschaft, "Multi-Scale Modeling from First Principles"
In this talk, Prof. Scheffler stressed that atomic level detail remains important even over long time scales. Examples focused on heterogeneous catalysis on 4d transition metal surfaces (Ru, Rh, Pd, Ag), all of which form an oxide film of varied thickness. In the presence of a simulated "constrained equilibrium" of an O2/CO gas mixture, ab initio atomistic thermodynamics allowed for the exploration of the 26 possible oxygen and carbon surface processes. Scheffler noted that over time scales possible with molecular dynamics, the surface remains relatively unchanged. Time scales on the order of tenths of seconds, possible with a kinetic Monte Carlo approach, are necessary to achieve a steady state condition where catalytic activity is prevalent. A rate equation approach, Phenomenological Kinetics, was also discussed which allows for the tracking of the long time evolution of mean field average surface site occupations.
• S. Savrasov, University of California Davis, "Computational Approaches for Strongly Correlated Materials: an Electronic Structure Theory Perspective"
This talk discussed the materials cases for which DFT fails, including high temperature superconductors, materials exhibiting colossal magneto-resistance, elements with partially filled f or d orbitals, and some organic materials. In these cases, Savrasov showed that DFT calculations incorrectly predict that insulators should be metallic above the magnetic ordering temperature and that narrow f bands are pinned at the Fermi level exhibiting no multiplet physics. For these cases, an alternative to DFT is necessary. Savrosov described a new approach in detail, spectral density functional theory, "which considers total free energy as a functional of a local electronic Green function."
• J.B. Neaton, Lawrence Berkeley National Laboratory, "First-Principles Studies of Electrical Transport in Nanoscale Molecular Junctions"
This talk addressed the comparison between experimental and simulated conductance measurements across a single molecule junction between two electrodes. The example presented was an amine molecule connecting Au electrodes. The HOMO-LUMO gap was used to determine conductance. DFT calculations resulted in values much higher than experiment which was determined to be due to not capturing the effect of the "image charge". A many-electron correction to DFT to account for this was discussed and shown to bring the computational results in line with experiment.
• V. Ozolins, University of California, Los Angeles, "First-Principles Thermodynamics and Kinetics of Advanced Hydrogen Storage Materials"
Prof. Ozolins discussed thermodynamic determining factors in choosing solid state materials for hydrogen storage that balance high density storage with relative ease of hydrogen release. Pertinent questions raised include: 1) What are all of the possible thermodynamic reversible reactions in a given multicomponent system? 2) What are the hydride crystal structures? 3) What are the rate limits for hydrogen absorption and desorption? DFT calculations allowed for the exploration of these questions, particularly for metal hydrides. Combined experimental and computational results for NaAlH4 were presented, a material for which desorption occurs at a relatively low 100°C.
• D. Srolovitz, Yeshiva University, "Liquid Metal Embrittlement: New Understanding for an Old Problem"
This talk addressed the degradation of ductile metals (Al, Cu, Ni, steel) when in contact with liquid metals (Ga, Hg, Pb, Bi, Li, Sn, Zn), a situation prevalent in processing. Prof. Srolovitz outlined factors governing liquid metal embrittlement (LME). First, LME typically occurs between metal pairs that do not form intermetallics. Second, LME occurs over a limited temperature range. Third, the effect increases with increasing grain size. Finally, it only occurs in the presence of a small applied stress. Molecular dynamics simulations of aluminum in contact with liquid gallium indicated that the applied stress is a critical factor, both for keeping cracks open and for causing dislocations to form, climb at a fixed rate, and allow for time independent Ga penetration. The host aluminum diffusivity at the grain boundaries also played a major role.
