[lammps-users] Error: Pair coefficient for Hybrid has invalid style

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps users,

I am trying to simulate a liquid on a surface.
The surface atoms ( 7 different types, bear charges) are interacting by L-J
potential only. It has only one type of bond O-H, other atoms are not connected.
The liquid atoms are all connected (molecules) and have bonds, angles,
dihedrals, vdw and coulomb interactions.

When I put only the surface or the liquid the simulation runs fine.( with
pair_style lj/cut/coul/ 25.0 25.0 ).
But when I put liquid on the surface with this pair style, the surface atoms
gets all mixed, and fly away, surface structure is lost.

So I thought I should use pair style hybrid:
But it is giving me the error : Pair coefficient hybrid has invalid style.

Here is my input :
clear
units real
#Defining the different potential types
pair_style hybrid lj/cut/coul/cut 20.0 lj/cut 2.5
atom_style full
angle_style harmonic
bond_style harmonic
dihedral_style harmonic
improper_style cvff

read_data clayffmod3bigwIL.lammps05

pair_coeff 1 1 lj/cut 0.0999999953 3.7423 2.5
pair_coeff 2 2 lj/cut 0.1554000006 3.5532 2.5
pair_coeff 3 3 lj/cut 0.1554000006 3.5532 2.5
pair_coeff 4 4 lj/cut 0.0000013298 4.7943 2.5
pair_coeff 5 5 lj/cut 0.0000018420 3.7064 2.5
pair_coeff 6 6 lj/cut 0.0000000000 0.0000 2.5
pair_coeff 7 7 lj/cut 0.0000018405 3.7064 2.5
pair_coeff 8 8 lj/cut 0.1554164124 3.5532 2.5

pair_coeff 9 9 lj/cut/coul/cut 0.1479999981 3.6170487995 20.0
pair_coeff 10 10 lj/cut/coul/cut 0.1669999743 3.5012320066 20.0
pair_coeff 11 11 lj/cut/coul/cut 0.1479999981 3.6170487995 20.0
pair_coeff 12 12 lj/cut/coul/cut 0.1669999743 3.5012320066 20.0
pair_coeff 13 13 lj/cut/coul/cut 0.0389999952 3.8754094636 20.0
pair_coeff 14 14 lj/cut/coul/cut 0.0380000011 2.4499714540 20.0
pair_coeff 15 15 lj/cut/coul/cut 0.0389999952 3.8754094636 20.0
pair_coeff 16 16 lj/cut/coul/cut 0.0950000000 3.5814000000 20.0
pair_coeff 17 17 lj/cut/coul/cut 0.0687685101 3.0808078941 20.0

pair_modify shift yes mix arithmetic

#Data Reading and neighbouring information
neighbor 2.0 bin
neigh_modify every 1 delay 0 check yes one 40000 page 2000000

#Minimization
min_style cg
dump 1 all xyz 100 dump_minclffsurf.xyz
minimize 0 1.0e-8 5000 5000
undump 1
write_restart restart_minclffsurf.300

Any suggestion is very much appreciated.

Thanks
Mousumi

2

Dear Lammps users,

I am trying to simulate a liquid on a surface.
The surface atoms ( 7 different types, bear charges) are interacting by L-J
potential only. It has only one type of bond O-H, other atoms are not connected.
The liquid atoms are all connected (molecules) and have bonds, angles,
dihedrals, vdw and coulomb interactions.

When I put only the surface or the liquid the simulation runs fine.( with
pair_style lj/cut/coul/ 25.0 25.0 ).
But when I put liquid on the surface with this pair style, the surface atoms
gets all mixed, and fly away, surface structure is lost.

so either you make a mess when you combine the
two systems, or the parameters for the mixed interactions
are not good.

So I thought I should use pair style hybrid:

i can't think of a reason why pair hybrid should
be fundamentally different, only that now you'll
have to provide all mixed parameter sets explicitly.

But it is giving me the error : Pair coefficient hybrid has invalid style.

do you have any pair coefficient definitions in your data file?

axel.

3

Hi,
Thank you very much for the reply.

I was trying pair_style hybrid because the manual says this style is good when
multiple pair potential is used in the simulation, for example metal ( eam
potential) on a surface( LJ potential).

I do not have any pair_style 'definition' for hybrid style on the data file ,
but I have the LJ paramaters listed in the data file.

I think I need to check the pair parameters.

Thanks
Mousumi

4

Hi,
Thank you very much for the reply.

I was trying pair_style hybrid because the manual says this style is good when
multiple pair potential is used in the simulation, for example metal ( eam
potential) on a surface( LJ potential).

yes, but those are very different functional forms and the simulations
would otherwise not be possible.

lj/cut is basically lj/cut/coul/cut with the charges set to zero
so they are not really different potentials.

axel.

5

I do not have any pair_style 'definition' for hybrid style on the data file ,
but I have the LJ paramaters listed in the data file.

That's probably why you're getting the error.
If you've set the pair_style to hybrid (when it used to be just lj/cut),
then those lines in your data file will have the wrong syntax. Since
you're setting all the pair coeffs in your input script, just remove
those lines from the data file.

Steve