I have the following situation. We have written a module for LAMMPS that is a modified SW interaction based on the paper by Oligschleger, C., et al., Model interatomic potential for simulations in selenium. Physical Review B (Condensed Matter), 1996. 53(10): p. 6165 that works quite well in LAMMPS and we have been running this for over a year with parameters for Te (yet to be published but in preparation). The module is copied from the SW pair_sw.cpp code and simply changes the force terms to match this modified form so that the rest of the code is identical to the SW module. We have new potentials for Te and Cd based on this form. Our potential form is referred to as hg in LAMMPS. It appears in the style list just fine and we have no other problems with it as far as I know. I am trying to run using the hybrid method as follows and obtain the ERROR below. Cd is type 1 atom and Te is type 2 atom. Does LAMMPS in the hybrid version check that the potential files incorporate 2 types of atoms? I cannot figure out the reason or meaning of this error. The corresponding potential files cd7.hg and te6.hg only contain one line corresponding to Cd Cd Cd or Te Te Te, respectively.
# Choose potential
pair_style hybrid sw hg
pair_coeff * * sw CdTe.sw Cd Te
pair_coeff 1 1 hg cd7.hg Cd
ERROR: Incorrect args for pair coefficients
Pair_coeff 2 2 hg te6.hg Te
ERROR: Incorrect args for pair coefficients
hi chuck,
I have the following situation. We have written a module for LAMMPS that is a modified SW interaction based on the paper by Oligschleger, C., et al., Model interatomic potential for simulations in selenium. Physical Review B (Condensed Matter), 1996. 53(10): p. 6165 that works quite well in LAMMPS and we have been running this for over a year with parameters for Te (yet to be published but in preparation). The module is copied from the SW pair_sw.cpp code and simply changes the force terms to match this modified form so that the rest of the code is identical to the SW module. We have new potentials for Te and Cd based on this form. Our potential form is referred to as hg in LAMMPS. It appears in the style list just fine and we have no other problems with it as far as I know. I am trying to run using the hybrid method as follows and obtain the ERROR below. Cd is type 1 atom and Te is type 2 atom. Does LAMMPS in the hybrid version check that the potential files incorporate 2 types of atoms? I cannot figure out the reason or meaning of this error. The corresponding potential files cd7.hg and te6.hg only contain one line corresponding to Cd Cd Cd or Te Te Te, respectively.
# Choose potential
pair_style hybrid sw hg
pair_coeff * * sw CdTe.sw Cd Te
pair_coeff 1 1 hg cd7.hg Cd
ERROR: Incorrect args for pair coefficients
Pair_coeff 2 2 hg te6.hg Te
ERROR: Incorrect args for pair coefficients
this was just discussed very recently on this list.
the problem here is that many-body potentials
_always_ have to have pair_coeff * *
and then you have to use NULL to have it
skip over the atoms that you don't want to
be treated by this specific potential.
e.g.:
pair_coeff * * sw CdTe.sw Cd NULL
pair_coeff * * hg te6.hg NULL Te
however, you _cannot_ have the self-self
interactions with one many-body potential
and the mixed interactions with another many-body
potential. this scenario only works, if the mixed
interactions are a pairwise potential and only makes
much sense, if you have two somewhat separated
phases, e.g. silicon and a metal.
to get what you seem to be after, you would have to
extend your custom pair style to include this all many-body
interactions into the same pair style. e.g. internally
select between your hg and the sw code.
cheers,
axel.