Berlin 2018 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
MM: Fachverband Metall- und Materialphysik
MM 67: Methods in Computational Materials Modelling (methodological aspects, numerics)
MM 67.6: Vortrag
Donnerstag, 15. März 2018, 18:45–19:00, TC 006
Effective treatment of formation energies for automated high-throughput computational materials design — •Rico Friedrich1, Cormac Toher1,2, Andrew Supka3,4, Marco Fornari1,3,4, Marco Buongiorno Nardelli1,5, and Stefano Curtarolo1,2 — 1Center for Materials Genomics, Duke Univ. — 2Mat. Sci., Elec. Eng., Phys. and Chem., Duke Univ. — 3Dept. of Phys., Central Michigan Univ. — 4Science of Advanced Materials Program, Central Michigan Univ. — 5Dept. of Phys. and Dept. of Chem., Univ. of North Texas
Automated high-throughput computational materials design, as implemented in the AFLOW framework [1], aims at the systematic prediction and optimization of materials properties for technological applications. The approach requires an accurate and efficient description of formation energies to assess the thermodynamic stability of new compounds. This presents a major challenge to standard computational materials science approaches such as density functional theory (DFT). Significant errors arise when calculating total energy differences between chemically dissimilar materials due to incomplete error cancellation [2]. This is particularly the case for oxides (and other
chalcogenides), where the chemical natures of diatomic (O2) molecules, elemental metals and solid oxides are very different. We discuss various approaches to compute the formation energies of materials and compare their accuracy with respect to experimental reference data.
S. Curtarolo et al., Comput. Mater. Sci. 58, 218 (2012).
V. Stevanović et al., Phys. Rev. B 85, 115104 (2012).