I am performing an all-atom molecular dynamic simulation in LAMMPS to perform the diffusion and adsorption capacities of gaseous hydrogen in metal alloys. For the simulation, I am using an embedded atom potential to model the interaction between the metal elements and a Lennard-Jones potential to model the H-H and H-metal elements interactions. All the lennard jones interactions follow the Lorentz-Berthelot mixing rules. I have used 0 as both the epsilon and sigma for the H-H interactions and by doing that all the other epsilons between the H-metal element are automatically going down to 0 (as the mixing rule for epsilon follows the geometric mean).
I am concerned about whether I am doing it right by using 0 LJ parameters for the H-H interaction. I have found some literature to support this for the adsorption but not for the case of diffusion analysis. It will be highly appreciated if anyone can give me some thoughts/advice on this.
pair_coeff * * eam/alloy 12_eam.alloy Al Cu NULL
pair_coeff 2 3 lj/cut/coul/cut 0 1.344
pair_coeff 1 3 lj/cut/coul/cut 0 1.1597
pair_coeff 3 3 lj/cut/coul/cut 0 0
Setting the LJ parameters of H2 to zero is not a good idea.
You did not specify the metals nor the temperature range. At low temperatures you need quantum corrections. Moreover, in some metals H2 dissolves under dissociation (Pd, Pt).
LJ pair potentials are generally too repulsive. They may, of course, work well if fitted for a specific application and temperature range.
Thanks for the answer, Prof. Ulrich Deiters. However, in this case, the metals I am using are Ni, Fe and Cr and the interactions are calculated at temperatures ranging from 900-1300K. Therefore, is it recommended to set the LJ parameters to zero here? Otherwise, some literature has used sigma = 2.915 Angstorm, epsilon = 0.316 kJ/mol for H adsorption with carbonaceous materials (Article Enhancement in hydrogen storage in carbon nanotubes under mo...
Article Molecular dynamics simulations of binary mixtures of methane...
), whereas others have used both as zero (Article Molecular Dynamics Simulation of Hydrogen Adsorption on Silica
)These are quite contradictory to each other which raised my doubts about what LJ parameters should I use for Hydrogen. I would appreciate any further insights into this matter.
The interaction of hydrogen with metals at such high temperatures is not in my field of expertise. But I can tell you this: If hydrogen dissolves in the metal as a molecule you need a non-zero sigma. If it dissolves by forming a metal hydride the H atoms will probably lie within the outer orbitals of the metal atoms, and therefore not require additional space. You need to find out first whether hydrogen dissolves in molecular form or undergoes a chemical reaction.
Incidentally, the Berthelot rule for estimating cross pair potentials is good for nonpolar, non-reacting molecules only, e.g., (H2 + CH4). It is not safe to assume its validity for the interaction of a hydrogen atom and a nickel atom.
- Badges
- Science topic