[lammps-users] Total and potential energies with EAM potentials

LAMMPS LAMMPS Mailing List Mirror
1

Problem:
I am seeing no variation in system potential (pe or pair or vdwl or bond)
energy when shearing a crystalline Al system in incremental amounts
perpendicular to the glide plane of an edge dislocation (slab geometry,
displace top plane of atoms parallel to glide plane, holding bottom plane
fixed, minimize and/or low-temp dynamics quenched to 0K, repeat). Atomic
motion appears correct (i.e. The dislocation is stationary at first then
starts to glide). I am looking at the potential energy difference (with
respect to the potential energy before strain is applied), so I expect to
see this difference increase as the system is elastically deformed (before
the dislocation moves) then remain more or less constant (small fluctuations
as dislocation glides). Because I am taking a difference, I expect this
measure should be sensitive to the small change in energy. Despite many
repeated attempts with different strain increments, cell sizes, geometries
etc. I see nothing but no/tiny random fluctuations in the potential energy.
I have tried using both eam potentials for Al provided with LAMMPS. I have
verified that the atoms do indeed move, and that the forces on the atoms
change.

I think the forces are correct (giving the correct atomic motions) but
perhaps the energies are not calculated/reported the way I expect.

Questions:
Which terms are included in the potential energies as reported by LAMMPS for
EAM potentials? Are there any missing components?

Is there any other trick or setting that immediately springs to mind that I
may have overlooked?

I am really stumped as to why the energy does not vary when the atomic
positions and forces change.

Any help would be appreciated greatly

Rebecca Powles
Electron Microscope Unit
University of Sydney, Australia

2

Questions:
Which terms are included in the potential energies as reported by LAMMPS for
EAM potentials? Are there any missing components?

Not clear if you mean total PE or pe/atom, but the EAM potential
reports all energy. The PE is consistent with the forces. If your system
is doing the right thing dynamically, then there should be no way the total
PE is wrong. If you're running NVE, you should conserve total E = KE + PE.
So any fluctuation in the temperature should be inversely fluctuating in the PE.

Is there any other trick or setting that immediately springs to mind that I
may have overlooked?

No.

I am really stumped as to why the energy does not vary when the atomic
positions and forces change.

Does your T change? If not, then the PE won't change much either.

Maybe what you want to do is print out per-atom PE and look at the energy
of atoms near the dislocation. If the dislocation moves, then the PE/atom
should also change in the neighborhood.

Steve