Researchers from the University of Helsinki, including Eetu Makkonen, Alvaro Lopez-Cazalilla, and Flyura Djurabekova, have developed a new computational tool to better understand and potentially mitigate hydrogen embrittlement in iron, a significant challenge for the energy industry, particularly in sectors like hydrogen storage and nuclear power.
Hydrogen embrittlement is a phenomenon where the presence of hydrogen atoms causes metals like iron to become brittle and prone to mechanical failure. This issue is particularly relevant to the energy sector, where iron and steel components are often exposed to hydrogen-rich environments. Understanding and predicting this behavior is crucial for designing safer and more durable materials for energy infrastructure.
The researchers have created a machine-learned interatomic potential for the alpha-iron-hydrogen system using the tabulated Gaussian Approximation Potential (tabGAP) formalism. This potential is trained on a dataset derived from Density Functional Theory (DFT) calculations, which provide highly accurate quantum mechanical descriptions of atomic interactions. The new potential is designed to accurately model the behavior of hydrogen atoms in iron, including their interactions with defects, dislocations, and other hydrogen atoms.
The tabGAP model outperforms existing classical and machine-learned potentials in predicting fundamental properties of the alpha-iron-hydrogen system. It achieves nearly DFT-level accuracy at a computational cost comparable to efficient classical Embedded Atom Method (EAM) potentials. This means it can provide detailed insights into hydrogen embrittlement without the extensive computational resources required for full DFT calculations.
The researchers demonstrated the utility of the tabGAP model by simulating tensile tests of iron structures with and without hydrogen atoms. The simulations revealed that hydrogen accelerates the decohesion of iron atoms at the tip of a notch, supporting the hypothesis of hydrogen-enhanced decohesion (HEDE) as a mechanism for hydrogen-induced mechanical failure. Additionally, the simulations showed an increase in vacancy concentration driven by hydrogen-dislocation interactions, supporting the hypothesis of hydrogen-enhanced strain induced vacancies (HESIV).
These findings are significant for the energy industry as they provide a more accurate and efficient tool for studying hydrogen embrittlement. By better understanding the mechanisms behind this phenomenon, engineers can design materials and structures that are more resistant to hydrogen-induced failure, leading to safer and more reliable energy infrastructure.
The research was published in the journal Computational Materials Science, providing a valuable resource for researchers and industry professionals working on materials for energy applications.
This article is based on research available at arXiv.