Regensburg 2019 – wissenschaftliches Programm
MM 4.1: Vortrag
Montag, 1. April 2019, 10:15–10:30, H45
Formation enthalpies for automated computational materials design — •Rico Friedrich1, Demet Usanmaz1, Corey Oses1, Andrew Supka2, Marco Fornari2, Marco Buongiorno Nardelli3, Cormac Toher1, and Stefano Curtarolo4 — 1Dept. of Mech. Eng. and Mat. Sci., Duke Univ. — 2Dept. of Phys. and Science of Advanced Materials Program, Central Michigan Univ. — 3Dept. of Phys. and Dept. of Chem., Univ. of North Texas — 4Mat. Sci., Elec. Eng., Phys. and Chem., Duke Univ.
The accurate calculation of formation enthalpies is crucial
for computational materials design. For compounds chemically similar to their reference phases such as metal alloys, standard semi-local approximations to density functional theory (DFT) lead to accurate results . When the
phases are chemically dissimilar as in the case of oxides, DFT suffers from a lack of error cancellation leading to deviations of several hundred meV/atom compared to experimental values . We use the automated computational
materials design framework AFLOW  to validate correction schemes for ab-initio formation enthalpies [2, 4]. These empirical methods can improve DFT predictions by a factor of 4 to 7. Zero-point vibrational and thermal
contributions to the formation enthalpy are found to largely cancel each other.
S. Curtarolo et al., Calphad 29, 163-211 (2005).
V. Stevanović et al., Phys. Rev. B 85, 115104 (2012).
S. Curtarolo et al., Comput. Mater. Sci. 58, 218 (2012).
L. Wang et al., Phys. Rev. B 73, 195107 (2006).