I try to introduce in LAMMPS, the tabulated potentials.
My input potential file is:
Pair potential for atom types 1 1 i,r,energy,force
Ge-Ge
N 1000 R 0.1 6
1 0.100000 6.752174 -2.43226859917
2 0.105906 1.433602 -1.3536564063
…
1000 0.1054506 2.435258 -9.3536545853
In my param file, I used this command bellow to introduce in my simulation, the potential from the file see previously.
Command in the param file:
pair_style table spline 1800
pair_coeff 1 1 Ge46_potentials_eopp_LAMMPS.pot Ge-Ge 6
rcutoff = 6 Angstrum.
There 1 type of atoms ‘Ge’.
1/ When I run the simulation with the number of point for the spline method > 1800 I find this error:
ERROR on proc 0: Pair distance < table inner cutoff (pair_table.cpp:112)
2/ But, if the number of point for the spline method < 1800, I have this error:
ERROR: Lost atoms: original 9936 current 866 (thermo.cpp:385)
2/ But, if the number of point for the spline method < 1800, I have this
error:
ERROR: Lost atoms: original 9936 current 866 (thermo.cpp:385)
Do you know the solution for my problem ?
there is not enough information here to give any specific advice.
the first error occurs due to atoms being too close. that can be due
to a bad initial structure or a bad potential tabulation.
the second error is more likely due to a bad table or the fact that
the potential tabulation will trigger artefacts from the spline
interpolation.
in general, tabulation requires significantly more care and testing
than analytical potentials, so it is strongly advised to gain
experience while practicing building tables and testing settings using
potentials that have an analytical counterpart in LAMMPS. for these it
is easily possible to determine the accuracy of the tabulation.
for further debugging, you can also write out the table at higher
resolution using pair_write and use that to test for spline artifacts.
the table pair style has multiple options to adapt to the specific
needs of a particular interaction. for example, in many cases, it is
advantageous to have a higher resolution at short distances, i.e. use
RSQ type tables, since potential functions tend to be steeper and more
curved at short r.