EAM Potentials

We developed two new embedded atom (EAM) potentials for hydrogen in BCC iron. The fitting targets (energies and atomic configurations) are generated from density functional theory (DFT) calculations for several equilibrium and non-equilibrium configurations. Initial tests indicate that these potentials perform significantly better than existing Fe-H potentials. (We would appreciate hearing from you if your tests reveal otherwise!) The Fe parts of the two potentials are taken from Mendelev et al. (Potential 4 in Philos Mag 11, 3977 (2003)) and Ackland et al. (J Phys: Condens Mat 16, S2629 (2004)). Both these Fe potentials produce the same screw dislocation core structure as DFT calculations, unlike most other Fe potentials. Our Fe-H potentials inherit this feature and should certainly be more suitable for modeling the interaction of H with dislocations. If you want to get started quickly, we recommend using Potential B.

(These EAM potentials were modified by Jun Song and Bill Curtin to avoid clustering of H atoms over extended time scales — additional details can be found in their paper. The potential files provided here DO NOT include Song and Curtin’s modifications.)

Downloads

Potential files generated in our group can be found on github: https://github.com/EACcodes/EAMpotentials

Please contact Prof. Ashwin Ramasubramaniam (former postdoc) directly for the source code  (C program) for tabulating the Fe-H EAM potential (Potential B) in LAMMPS format.

These potential files are compatible with LAMMPS (as of Dec 2008).

Preferred potentials:
We recommend using Potential B, which performs slightly better than Potential A in our tests. You must run these with “pairstyle = eam/fs” (see LAMMPS documentation for details).

Potential A : This is based on Fe Potential 4 from Mendelev et al. (Philos Mag 11, 3977 (2003)) and fit to bulk, vacancy, and surface data from DFT.

Potential B : This is based on the Fe potential from Ackland et al. (J Phys: Condens Mat 16, S2629 (2004)) and fit to bulk, vacancy, and surface data from DFT.

(Note: These potentials were revised on February 24, 2010. If you downloaded prior versions, we strongly recommend upgrading to the latest ones.)

“Bulk-only” potentials:
These potentials were fit only to bulk and vacancy DFT data and are slightly better than Potentials A and B in their response to strain fields in bulk environments. However, these do not perform well at surfaces. Use at your own discretion! Also note that you must run these with “pairstyle = eam/alloy” (see LAMMPS documentation for details).

Potential A’ : This is based on Fe Potential 4 from Mendelev et al. (Philos Mag 11, 3977 (2003)) and fit only to bulk and vacancy data from DFT.

Potential B’ : This is based on the Fe potential from Ackland et al. (J Phys: Condens Mat 16, S2629 (2004)) and fit only to bulk and vacancy data from DFT.