BOX EXPANSION DURING EQUILIBRATION

LAMMPS LAMMPS Mailing List Mirror
1

Dear users,

I have modeled a cnt through vmd with a molecular atom style(bonds,angles, dihedrals included).
I am using periodic boundary conditions and a box that is quite larger(10 times) than the dimensions of the cnt in x,y directions and equal to the dimension of cnt in z direction.
During equilibration i use an npt fix 300 300 1 iso 0 0 1 drag 1
I run it for 400000 time steps and although the potential energy equilibrates and the temperature somewhat does the box is expanding 5 times its size and does not stabilize.
I also get a warning : bond/angle/dihedral extent >half of periodic box size.
If you any idea about the cause of the problem or any suggestion regarding its fix please let me know.

Thank you very much,
D.G

2

Dear users,

I have modeled a cnt through vmd with a molecular atom style(bonds,angles,
dihedrals included).
I am using periodic boundary conditions and a box that is quite larger(10
times) than the dimensions of the cnt in x,y directions and equal to the
dimension of cnt in z direction.

​why so large in x and y? why not using fixed boundaries there?
why so tight in z? ​

During equilibration i use an npt fix 300 300 1 iso 0 0 1 drag 1

​why couple the the box size in x and y to the box size in z and only allow
isotropic box changes?
why use fix npt in the first place? ​

I run it for 400000 time steps and although the potential energy
equilibrates and the temperature somewhat does the box is expanding 5 times
its size and does not stabilize.
I also get a warning : bond/angle/dihedral extent >half of periodic box
size.
If you any idea about the cause of the problem or any suggestion regarding
its fix please let me know.

​well, do you have one or more bonds with that extreme length?​ e.g. by
adding bonds across periodic boundaries.

it is difficult to make any specific suggestions, since you don't provide
any specific information, but only rather vague descriptions and no
explanations for rather unexpected choices.

axel.

3

Hello Dr Axel,
I am using these specific box dimensions because they were created like this by vmd when i wrote the file in the lammps input form.
I have tried using a box that matches the dimensions of the nanotube but i get the same results. The box increases in size in a large amount.
I am using npt in order to equilibrate my system at zero pressure and temperature 300 K so i can later compute the young modulus with a strain rate of 0.01.
This is the script i am using :

units metal
dimension 3
boundary p p p
newton on
atom_style molecular

----------------------- ATOM DEFINITION ----------------------------

read_data cnttopology

------------------------ FORCE FIELDS ------------------------------

pair_style airebo 3
pair_coeff * * …/potentials/CH.airebo C

------------------------- SETTINGS ---------------------------------

compute 1 all ke
compute 0 all pair airebo
variable REBO equal c_0[1]
variable LJ equal c_0[2]
variable TORSION equal c_0[3]

4

there were many more, important questions i have asked that are still unanswered.
…and here is one more: why do you use atom style molecular, even though you are using airebo, where bonds/angles/dihedrals are implicit?
the error message about long bonds suggests, that you have bonds defined in your topology. why?

axel.