[lammps-users] remapping atoms during minimization using fix box/relax

Dear Steve,
I have the following problem. I am trying to minimize a CNT structure with fix box/rescale option so that the final pressure of the system is close to zero. However, when I looked at the forces on the atom (in the minimized structure), the forces on the atoms which are close to the boundary (I am using periodic boundary condition) is much higher then the forces on the atoms in the centre of the system. These forces have both positive and negative signs (high fluctuations) and hence cancel out resulting in zero pressure, I think. Based on this, it seems to me that the fix box/rescale option just changes the dimension of the box, but does not rescale the coordinates, is it possible to include “remap x” option in fix box/relax command so that the pressure is ditributed evenly along the box.

The reason I am asking for this (instead of fix deform command) is that we are planning to series of molecular mechanics simulations (not dynamics), so fix deform would not be helpful.

I could use displace_box also to remap the coordinates, however, the problem with it is that is does the rescaling at once. However, box/rescale rescales the box at each minimization step. I can’t send attachments (as a plot) from work through gmail. So, I am writing the values of the forces (relative) below for minimized structure.

box dimension along Z xlo=22.1579 xhi=2451.29. this is 6,6 nanotube so there are 12 atoms along each ring.

Atom No. Force value (eV/Angs) Z value
1-12 7.97e-2 22.7666 (they have same Z coodinate, very close to x low value)
12-24 5.14e-3 23.9811 (they have same Z coodinate)
25-23976 0
23977-23988 -5.14e-3 2449.4718 (they have same Z coodinate)
23989-24000 -7.97e-3 2450.6864 (they have same Z coodinate)

Any help is appreciated.

Many regards,
Vikas Varshney

I assume you mean the new fix box/relax, not box/rescale?
Aidan, do you want to comment on this?

Steve