[lammps-users] why graphene break up after 1ps simulation with AI-REBO potential

Hi, all,

I’m simulating a small graphene ribbon of 96 carbon atoms with pbc in xyz directions. I use AI-REBO potential. I use the following data and input files to begin lammps simulation. After 10ps simulation, I check the trajectory file and find the graphene ribbon breaking up in 1ps. So can you give me some help or tell me where the errors are in my data and input files?

Thanks so much!

Dongshan

You should use 'metal' units instead of 'real' units.

Zhun-Yong

Why? Whether or not is there a corresponding relation between potentials and units? If I use reaxff potential, what units should I use? What are the energy units for parameters in ffield.reax and CH.airebo potential files?

Thanks!

Dongshan

Why? Whether or not is there a corresponding relation between potentials and

please read the documentation. the AI-REBO potential files are
in metal units and the pair style cannot convert them to something else.

units? If I use reaxff potential, what units should I use? What are the

RTFM.

energy units for parameters in ffield.reax and CH.airebo potential files?

in principle the pair style can be written to convert as
needed, but people are lazy. feel free to submit a patch
that makes up for it. i am certain that a significant number
of (now confused) LAMMPS users will appreciate it.

cheers,
   axel.

Hi Axel,

Thanks for your reply. I make clear the units of different potentials. But in my case, when I change to use the metal unit, the graphene sheet breaks up faster. At t=0.5ps, the system temperature increases to 1000K and the sheet breaks up. I think there must be problems in my input file or dta file. Can you help me check my input file and give your modification?

Thanks!

Dongshan

PS: the modified input file is as follows:

Hi Axel,

Thanks for your reply. I make clear the units of different potentials. But
in my case, when I change to use the metal unit, the graphene sheet breaks
up faster. At t=0.5ps, the system temperature increases to 1000K and the
sheet breaks up. I think there must be problems in my input file or dta
file. Can you help me check my input file and give your modification?

you cannot use a time step of 0.5 ps for carbon atoms. they are far too
light for that. use 0.0005 and it should work much better. note that the
default unit for time step is "ps" in "metal" and "fs" in "real" units.

what you see is just the behavior of a diverging integrator.

cheers,
    axel.

You are also missing an 'unfix 1'. Didn't you get an error message
warning you of multiple fixes when you tried running the input script?

Zhun-Yong

The doc pages for both pair styles tell you what units the potential
files are in.

Steve