Toward machine learning for microscopic mechanisms: A formula search for crystal structure stability based on atomic properties

Abstract

Machine-learning techniques are revolutionizing the way to perform efficient materials modeling. We here propose a combinatorial machine-learning approach to obtain physical formulas based on simple and easily accessible ingredients, such as atomic properties. The latter are used to build materials features that are finally employed, through linear regression, to predict the energetic stability of semiconducting binary compounds with respect to zinc blende and rocksalt crystal structures. The adopted models are trained using a dataset built from first-principles calculations. Our results show that already one-dimensional (1D) formulas well describe the energetics; a simple grid-search optimization of the automatically obtained 1D-formulas enhances the prediction performance at a very small computational cost. In addition, our approach allows one to highlight the role of the different atomic properties involved in the formulas. The computed formulas clearly indicate that “spatial” atomic properties (i.e., radii indicating maximum probability densities for [Formula: see text] electronic shells) drive the stabilization of one crystal structure with respect to the other, suggesting the major relevance of the radius associated with the [Formula: see text]-shell of the cation species.

Recent Advances in the Application of Machine Learning to Crystal Behavior and Crystallization Process Control

“…In addition to melting point and crystal habit prediction, machine learning has been explored in the prediction of a number of other crystal properties. Some results have been achieved in predicting the stability of crystal structures. − Using only two descriptors, namely Pauling electronegativity and ionic radius, Ye et al can predict the density functional generation energy of C 3 A 2 D 3 O 12 garnet and ABO 3 perovskite with an average absolute error of 7–10 meV/atom and 20–34 meV/atom, respectively, which are within the limits of density functional accuracy. A critical gap in the extension of machine learning models from fixed stoichiometric ratio crystals to the realm of infinite mixture crystal systems has been addressed.…”

Understanding the Magnetic Microstructure through Experiments and Machine Learning Algorithms

“…The ions lose energy during their passage through the material, which is either spent in displacing the target atoms by elastic collision (nuclear stopping) or exciting the atoms by inelastic collisions (electronic stopping) . A large variety of studies exists on ion-beam induced modifications in magnetic ML, either in the context of patterning or depth-resolved structural modifications. , The creation of graded anisotropy media by domain wall positioning has also been reported using ion irradiation . Ion-induced modification of magnetic properties with depth-resolved structural studies in Co/Pt ML has been reported. , An interesting study on the investigation of magnetic domains after ion beam irradiation was performed by Trassinelli et al, where local microscopic features of domains were highlighted in the vicinity of the ferromagnetic to paramagnetic phase transition temperature for Mn–As thin films .…”

“…28 A large variety of studies exists on ion-beam induced modifications in magnetic ML, either in the context of patterning 29 or depth-resolved structural modifications. 30 , 31 The creation of graded anisotropy media by domain wall positioning has also been reported using ion irradiation. 32 Ion-induced modification of magnetic properties with depth-resolved structural studies in Co/Pt ML has been reported.…”