Fast machine learned $α$-Fe-H interatomic potential for hydrogen embrittlement

TL;DR

Hydrogen embrittlement in the $\ ext{alpha}$-Fe–H system arises from complex defect interactions that are not well captured by simple potentials. The authors develop a tabGAP-based interatomic potential for $\ ext{alpha}$-Fe–H trained on DFT data, combining 2-body, 3-body, and EAM descriptors with a short-range repulsion to reproduce H interactions at defects. They validate the model against DFT across H point defects, vacancies, dislocations, H–H interactions, and elastic moduli, and demonstrate HE mechanisms HEDE and HESIV through MD tensile simulations. The tabGAP achieves near-DFT accuracy with computational efficiency close to classical EAM, enabling large-scale simulations of hydrogen embrittlement in iron under realistic conditions.

Abstract

In this work, we present a machine-learned interatomic potential for the $α$-Fe-H system based on the tabulated Gaussian Approximation Potential (tabGAP) formalism. Trained on a Density Functional Theory (DFT) dataset of atomic configurations, energies, forces, and virials, the potential is designed to address the issue of H-induced acceleration of mechanical failure of metals, generally known as hydrogen embrittlement (HE). The proposed potential is shown to outperform the widely used classical and machine-learned interatomic potentials in fundamental properties of the $α$-Fe-H system. We show that the tabGAP model reproduces H-point defect properties, H-dislocation interaction, H-H interaction, and elastic constants with nearly DFT-level accuracy at a computational cost that is competitive with the efficient classical Embedded Atom Method (EAM) potentials. We further demonstrate the utility of the tabGAP model in molecular dynamics simulations of tensile tests of a perfect, and a $(111)[11\bar{2}]$-notched $α$-Fe structure with and without the load of H atoms. The simulations show evidence of HE via observation of accelerated decohesion of Fe atoms at the tip of the notch, and an increase in vacancy concentration driven by H-dislocation interactions. Hence, the results of the presented simulations support the hypotheses of hydrogen-enhanced decohesion (HEDE) alongside hydrogen-enhanced strain induced vacancies (HESIV) as important mechanisms for H-induced mechanical failure of iron systems.

Figures (16)

• Figure 1: The migration energy barriers of the T-T (\ref{['fig:TT_paths']}) and T-O-T transitions (\ref{['fig:TOT_paths']}) for a H atom in bulk -Fe, with sketches provided for both. DFT data is from Ref. haywardInterplayHydrogenVacancies2013b. The sketches show a [100]-view of the bcc unit cell. The H atoms are colored red, and Fe atoms are gray. DFT data is taken from Ref. haywardInterplayHydrogenVacancies2013b.
• Figure 2: The binding sites and energies for a H atom near a $1/2\hkl<111>\hkl{110}$ soft-core screw dislocation, calculated with the tabGAP (Fig. \ref{['fig:H_Eb_screw_pos']}), and the binding energies of a few annotated cites compared to DFT calculations done by Itakura et al.itakura_effect_2013 (Fig. \ref{['fig:H_Eb_screw_dft']}).
• Figure 3: The binding sites near a $1/2\hkl<111>\hkl{110}$ edge dislocation for a H atom, and their binding energies, calculated with the tabGAP.
• Figure 4: The binding energy of interstitial H-pairs within the -Fe matrix calculated with the Eq. \ref{['eq:Eb_HH']}. The x-axis is the distance (fractional units) between the pair of H atoms after energy minimization. DFT results are from Ref. haywardInterplayHydrogenVacancies2013b.
• Figure 5: The radial distribution functions of Fe-H and H-H pair separation distances in a ($37\times37\times37$)-sized -Fe $10\atp$-H-structure before and after equilibrating for $1ns$ at $300K$ with the NPT ensemble.