Dresden 2017 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 55: Topical session: Data driven materials design - structure maps

MM 55.5: Talk

Thursday, March 23, 2017, 11:15–11:30, BAR 205

Compressed-Sensing Models for the Prediction of the (Meta-)Stability of Octet Binaries — •Emre Ahmetcik, Runhai Ouyang, Christian Carbogno, Luca M. Ghiringhelli, and Matthias Scheffler — Fritz-Haber-Institut der MPG, Berlin

Predicting the ground-state and metastable crystal structures of materials from just knowing their composition is perhaps the most important challenge in materials science. In a recent showcase study [1], a compressed-sensing based methodology was introduced to identify the descriptors (nonlinear functions of atomic input features) that best describe and predict the relative stability of zincblende versus rocksalt octet binary compounds. In this work, we have extended this approach to describe the relative stability and energy differences between eight different polymorphs of the octet binary materials, so that an accurate prediction of their ground and metastable states becomes possible. We discuss how this is tackled by introducing a new class of descriptors that map the geometry of different structures into a well defined scalar (the “reduced” Madelung energy) and a multi-objective optimization (the simultaneous minimization of the prediction error of all formation-energy differences among the eight considered polymorphs).
This work received funding from The Novel Materials Discovery (NOMAD) Laboratory, a European Centre of Excellence.

[1] Ghiringhelli et al., PRL 114, 105503 (2015).