[lammps-users] Some troubles with EAM potential

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS users
The question about units in EAM potential files was here already
(http://sourceforge.net/mailarchive/forum.php?thread_id=26117938&forum_id=46682).
But I met same problem as in that message. When I try my own potential
file I obtain very large values of energy and pressure. To find origin of
the problem I performed simple test.
At first I wrote simple potential file:

Test potential
47 107.868 1.00 FCC
20 1.0e+00 10 0.1e+00 0.9e+00
    0.0e+00 1.0e+00 2.0e+00 3.0e+00 4.0e+00 5.0e+00 6.0e+00 7.0e+00
8.0e+00 9.0e+00 10.0e+00 11.0e+00 12.0e+00 13.0e+00 14.0e+00
15.0e+00 16.0e+00 17.0e+00 18.0e+00 19.0e+00
   0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00
0.0e+00 0.0e+00
   0.0e+00 0.1e+00 0.2e+00 0.3e+00 0.4e+00 0.5e+00 0.6e+00 0.7e+00
0.8e+00 0.9e+00

There are only attractive part of potential in this file. Then I
calculated energy per atom for FCC lattice with lattice parameter 1.0. It
was OK I obtained expected value -8.48 eV

Then I change this potential file to put there only repulsive forces:

Test potential
47 107.868 1.00 FCC
20 1.0e+00 10 0.1e+00 0.9e+00
    0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00
0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00
0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00
   1.0e+00 1.0e+00 1.0e+00 1.0e+00 1.0e+00 1.0e+00 1.0e+00 1.0e+00
1.0e+00 1.0e+00
   0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00 0.0e+00
0.0e+00 0.0e+00
I expected to get energy per atom as 6 eV (0.5*12*fi - where 12 is number
of neigbours) but I obtained 122.09 eV. It is certainly very large value
and certainly it results in large pressure. So I can conclude that
something wrong with repulsive forces in potential file but I don't
understand what exactly.
I will appreciate for answers.
Sincerely yours
Andrey Ostapovets

2

All I can suggest is you plot your fitted functions (rho, F, phi)
versus the ones in an existing potential file to see what shape
they have and how they work. You can also use the pair table
command in LAMMPS to plot out the interpolated phi function
at any resolution you want, to see if it is what you expect. For
EAM, you can only do this for the pair potential part of the EAM
potential.

Steve

3

Andrey,
The pair potential in the EAM file is in different formats for "funcfl"
and "setfl" files.
In funcl files, it is written as sqrt(r*phi(r)).
Setfl files have it written as r*phi(r).

Hence the values go from r = 0 to r = rcut-dr. (This is how LAMMPS will
assume the values are written.)

* NOTE: dr should always be rcut/nr.

You can find further details here :
http://web.ics.purdue.edu/~asubrama/pages/EAM_Potential_format.htm

I have checked that the energy is correctly calculated by using a simple
LJ potential. If you want I can send you the matlab code I used to write
the EAM potential file. Do let me know if that would be helpful.

Hope this helps,
Arun

Andriy Ostapovets wrote: