Oxidation-dependent mechanical response of graphene oxide: Improving the reliability of atomistic modeling by ab initio machine learning simulations
Physical Review Materials · 02.11.2026 · 10 (11)
DOI: 10.1103/vwr6-mt41 ↗Kurz gefasst
Fachartikel in der Zeitschrift Physical Review Materials, online erschienen am 02.11.2026. Verfasst von Sara Shahbazi Fashtali und 2 weiteren Autor:innen. Die Arbeit nennt 69 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NextGenerationEU, Donostia International Physics Center, Barcelona Supercomputing Center.
Abstract (Originalsprache)
The mechanical response of graphene oxide (GO) is controlled by oxidation level and functional group composition, yet a clear atomic level picture remains ambiguous due to limitations of classical reactive force fields. Here, we develop a machine learning interatomic potential by fine-tuning a MACE foundation model on density functional theory data. We employ this potential to systematically investigate GO tensile behavior across oxidation levels of 10% to 40% and hydroxyl-to-epoxy (OH/O) ratios of 0 to 3. Our potential accurately reproduces the expected decrease in Young's modulus and ultimate tensile strength with increasing oxidation, correcting the overestimation and dispersion characteristic of ReaxFF predictions. In the elastic and plastic regimes, oxidation level dominates the mechanical response, while the functional group identity remains secondary. In the fracture regime, however, chemical composition becomes significant, controlling ductile/brittle behavior. In highly oxidized, hydroxyl-rich systems, stress relaxation via dehydration, not captured by reactive force fields, critically influences fracture behavior.
Förderung
- NextGenerationEU
- Donostia International Physics Center
- Barcelona Supercomputing Center
- European Social Fund Plus
Zitieren
APA
Fashtali, S. S., Piaggi, P. M., & Zollo, G. (2026). Oxidation-dependent mechanical response of graphene oxide: Improving the reliability of atomistic modeling by ab initio machine learning simulations. Physical Review Materials, 10(11). https://doi.org/10.1103/vwr6-mt41
Vancouver
Fashtali SS, Piaggi PM, Zollo G. Oxidation-dependent mechanical response of graphene oxide: Improving the reliability of atomistic modeling by ab initio machine learning simulations. Physical Review Materials. 2026;10(11). doi:10.1103/vwr6-mt41
BibTeX
@article{Fashtali2026,
author = {Fashtali, Sara Shahbazi and Piaggi, Pablo M. and Zollo, Giuseppe},
title = {Oxidation-dependent mechanical response of graphene oxide: Improving the reliability of atomistic modeling by ab initio machine learning simulations},
journal = {Physical Review Materials},
year = {2026},
volume = {10},
number = {11},
doi = {10.1103/vwr6-mt41}
}
EUVIMED