pair_style hybrid with eam/fs and lj

Dear LAMMPS users

Hi, I’m graduate student at Seoul National University in Seoul, Korea.

Currently I’m dealing with a metal and a organic compound mixture simulation.

I’m trying to use FS EAM for the metal and some class 1 type force field obtained from AMBER for the organic compound.

However, when I write the input script as

pair_style hybrid eam/fs lj/cut/coul/cut 5.0

pair_coeff 1 1 eam/fs Au.eam.fs Au

pair_coeff 2 2 lj/cut/coul/cut 0.0860 1.9080 #C

pair_coeff 3 3 lj/cut/coul/cut 0.0860 1.9080 #CC

…etc.

I get ‘incorrect args for pair coeff. …/pair_eam_fs’ error.

I’ve tried ‘pair_coeff * * eam/fs Au.eam.fs Au Null … N*Null’, of course this is not the one I want.

How can I exactly apply FS EAM for just type 1 and else as lj style?

Furthermore, I want to use ‘pair_modify mix arithmetic’ for the 2N 2N interactions.

(like 2-3, 2-4,…,2-N, 3-4,… interactions)

Will this also assign 1-N interactions?

I’m confused because the LAMMPS document for EAM style says ‘pair_coeff 2 12 cuu3.eam’ will ignore 1,2 and 2,1.

I understand that EAM is many body potential so that this is not legit.

I guess I should find some lj parameters for type 1 and apply Lorentz-Berthelot mixing rules for cross interactions.

Any suggestion will be greatly helpful.

Best regards,

Hyun-seok Kim.

Dear LAMMPS users

Hi, I'm graduate student at Seoul National University in Seoul, Korea.

Currently I'm dealing with a metal and a organic compound mixture
simulation.

I'm trying to use FS EAM for the metal and some class 1 type force field
obtained from AMBER for the organic compound.

However, when I write the input script as

pair_style hybrid eam/fs lj/cut/coul/cut 5.0

pair_coeff 1 1 eam/fs Au.eam.fs Au

pair_coeff 2 2 lj/cut/coul/cut 0.0860 1.9080 #C

pair_coeff 3 3 lj/cut/coul/cut 0.0860 1.9080 #CC

...etc.

I get 'incorrect args for pair coeff. ../pair_eam_fs' error.

​and rightfully so.​

I've tried 'pair_coeff * * eam/fs Au.eam.fs Au Null .... N*Null', of
course this is not the one I want.

​why not? that is the one you *have* to use. please see the eam/fs docs.​

How can I exactly apply FS EAM for just type 1 and else as lj style?

​using pair_style hybrid. you just have to do it right and follow the
documentation.

Furthermore, I want to use 'pair_modify mix arithmetic' for the 2*N 2*N

interactions.

(like 2-3, 2-4,...,2-N, 3-4,.... interactions)

Will this also assign 1-N interactions?

​no. the 1-1 interactions are of a completely different nature, thus there
is no way to extract​

I'm confused because the LAMMPS document for EAM style says 'pair_coeff
*2 1*2 cuu3.eam' will ignore 1,2 and 2,1.

​but that is for "eam" and *not* "eam/fs". it works, because "eam" does not
support alloys. ​with "eam/fs" alloys are possible and thus in addition to
'i-i' pairs all mixed pairs between types of "EAM" atoms have to be
considered as well. the '* *' notation enforces that.

I understand that EAM is many body potential so that this is not legit.

I guess I should find some lj parameters for type 1 and apply

Lorentz-Berthelot mixing rules for cross interactions.

​that may not be the best choice. any kind of LJ mixing​ rules result in
reasonable inter-type potentials only for a few cases. for it work, *all*
involved potentials have to be parameterized with the same parameterization
strategy for the same kind of environment. it often also requires that the
atoms are of similar kind of compounds (e.g. organic molecules). metals can
behave quite different, especially when put into contact with polar
compounds. so what is applicable depends a lot on how well you want to
represent the gold surface. you may be forced to use (in order of
decreasing level of detail) DFT, ReaxFF, a BOP, a jellium model or an image
charge model.

axel.