[lammps-users] Problem with pair_style hybrid

LAMMPS LAMMPS Mailing List Mirror
1

Steve,

I am simulating a system with carbon & silicon atoms. I want to use airebo for C-C and tersoff for Si-Si & Si-C atoms. I am getting error for the following part of the code.

1 is for silicon 2 for carbon here.

pair_style hybrid tersoff airebo 2.5 1 0
pair_coeff 1 1 tersoff SiC.tersoff Si NULL
pair_coeff 2 2 airebo CH.airebo NULL C
pair_coeff 1 2 tersoff SiC.tersoff Si C

Error: Incorrect args for pair coeff

I am not sure where the problem is.

Thanks
Ajit

2

Those potentials require a pair_coeff * * form
of the command. See the doc pages for Tersoff
and AIREBO.

Steve

3

Steve,
               If i use "* *" instead of no's , it is showing the message "segmentation fault".
            Can u please help me fix that ?

  pair_style hybrid tersoff airebo 2.5 1 0
    pair_coeff * * tersoff SiC.tersoff Si NULL
     pair_coeff * * airebo CH.airebo NULL C
     pair_coeff * * tersoff SiC.tersoff Si C

Thanks
Ajit

4

The seg fault was a memory bug with specifying the
tersoff coeffs twice. I'll post a patch. But you can't
do what you are writing below. You can't have AIREBO
operate on type 2 atoms, and Tersoff on type 1 and 2.
These are both many-body potentials. When Tersoff
acts on type 1 and 2, it will also act on interactions
between just type 2 atoms, overwriting the AIREBO
setting.

Steve

5

Steve,

                How about this?

  pair_style hybrid/overlay airebo 2.5 1 0 tersoff sw
    pair_coeff * * airebo CH.airebo NULL C
    pair_coeff * * sw Si.sw Si NULL
    pair_coeff * * tersoff SiC.tersoff Si C

    Does the above part ensure airebo for C-C, SW for Si-Si & Tersoff for C-Si??

Thanks
Ajit

6

Steve,

           How about this?

pair_style hybrid/overlay airebo 2.5 1 0 tersoff sw
pair_coeff * * airebo CH.airebo NULL C
pair_coeff * * sw Si.sw Si NULL
pair_coeff * * tersoff SiC.tersoff Si C

Does the above part ensure airebo for C-C, SW for Si-Si & Tersoff
for C-Si??

as steve already said this will do tersoff for _all_ atoms.
and then _add_ airebo for C-C and SW for Si-Si interactions.

overall, given the manybody nature of each of the
three potential types you are using, this whole idea of
treating them like pair-wise additive potentials does
not make any sense at all.

axel.