Poster De Conférence Année : 2026

Understanding oxygen-driven plasticity in amorphous alloys via multi-scale simulations and ab initio trained machine learning force fields

Résumé

Predicting the mechanical response of metallic glasses (MGs) remains challenging, as conventional atomistic models often rely on empirical force fields that fail to accurately capture complex deformation mechanisms or multicomponent chemical interactions. Here, we develop a multiscale modeling framework to investigate the brittle-to-ductile transition in ZrCu-based MGs, with particular emphasis on the role of oxygen concentration and structural state. A machine-learning force field (MLFF) is trained on ab initio density functional theory data and subsequently employed in large-scale molecular dynamics simulations to access system sizes and deformation regimes beyond the reach of first-principles methods. The MLFF is systematically validated by comparing ground-state properties and elastic constants with corresponding DFT results. Mechanical behavior is analyzed under compressive loading through stress–strain responses and the computation of avalanches. These results demonstrate that ML-based interatomic potentials provide a reliable framework for linking atomistic deformation mechanisms to macroscopic mechanical behavior in metallic glasses.

Fichier principal
Vignette du fichier
Slides.pdf (2.15 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
licence

Dates et versions

hal-05619853 , version 1 (12-05-2026)

Licence

Identifiants

  • HAL Id : hal-05619853 , version 1

Citer

R. Alvarez-Donado, Oguz Umut Salman, P. Djemia, Matteo Ghidelli. Understanding oxygen-driven plasticity in amorphous alloys via multi-scale simulations and ab initio trained machine learning force fields. MecaNano General Meeting 2026, Apr 2026, Rome, Italy. ⟨hal-05619853⟩
0 Consultations
0 Téléchargements

Partager

  • More