[lammps-users] tersoff potential problem

Dear All,

I am trying to define tersoff potential for 4 atoms but met some problem. There are 4 atom types in my simulation and I wanna make tersoff potential between 1 and 2, 3 and 4 (no tersoff potentail between 1 and 3, 1 and 4, 2 and 3, 2 and 4). In order to achieve it, I use following commands.

pair_style tersoff
pair_coeff * * SiC.tersoff C C NULL NULL
pair_coeff * * SiC.tersoff NULL NULL C C

which is not correct anyway. Any suggestion is greatly appreciated.

Regard,
ML

Dear All,

I am trying to define tersoff potential for 4 atoms but met some problem.
There are 4 atom types in my simulation and I wanna make tersoff potential
between 1 and 2, 3 and 4 (no tersoff potentail between 1 and 3, 1 and 4, 2
and 3, 2 and 4). In order to achieve it, I use following commands.

pair_style tersoff
pair_coeff * * SiC.tersoff C C NULL NULL
pair_coeff * * SiC.tersoff NULL NULL C C

which is not correct anyway. Any suggestion is greatly appreciated.

this cannot be done with lammps and it doesn't make
much sense either. please see previous discussions
on this subject, ie using the same many-body potential
type multiple times with pair style hybrid, in the mailing
list archives.

cheers,
    axel.

Axel is correct. The only thing you can specify
with a command like this:

pair_coeff * * SiC.tersoff C C NULL NULL

is that all the atoms of type 1 and 2 interact
between themselves via a Tersoff potential. If you
want any of those atoms to interact with some other
type (e.g. 3 or 4) then they cannot interact with those
atoms via Tersoff. That is what wouldn't make sense
for a many-body potential; it's not a well-defined interaction.

Steve