Question for net charge being not zero after deleting overlap atoms

LAMMPS
1

Dear all,

I built a LLZO grain boundary model using Atomsk. Because there were overlapping atoms at the GB region, I used the remove doubles command in Atomsk to delete atoms that were too close. After this removal, the stoichiometry of the system is no longer conserved, and the net charge becomes non-zero.The picture is Zr of my model.

image
image1600×1200 165 KB

In my LAMMPS input, I set:

text

boundary        p p p
atom_style      full
kspace_style    ewald 1.0e-5
pair_style      buck/coul/long 12
read_data       gb_rmd.data

set type 1 charge +1
set type 2 charge -1.65
set type 3 charge +2.50
set type 4 charge +2.65

pair_coeff 1 1  0 0.1 0
pair_coeff 2 2   4870.00 0.267 77.00
pair_coeff 3 3 0 0.1 0
pair_coeff 4 4 0 0.1 0
pair_coeff 1 2  1087.29 0.260 0       
pair_coeff 2 3   2075.26 0.326 23.25
pair_coeff 2 4  1650.32 0.311 5.10
pair_coeff 1 3  0 0.1 0
pair_coeff 1 3  0 0.1 0
pair_coeff 1 4  0 0.1 0
pair_coeff 3 4  0 0.1 0

When running LAMMPS, I got the warning:

text

  5376 settings made for charge
Setting atom values ...
  9152 settings made for charge
Setting atom values ...
  2246 settings made for charge
Setting atom values ...
  1476 settings made for charge
5376 atoms in group Li
Ewald initialization ...
WARNING: System is not charge neutral, net charge = -198.4

The simulation can still proceed, and I plan to calculate the local diffusivity via MSD.

My questions are:

  1. How serious is this net charge problem? Will it significantly affect the simulation results (especially MSD and diffusion coefficients)?

  2. Can I simply ignore this warning and continue, or must I correct the net charge to zero?

  3. If I need to fix it, would it be acceptable to randomly add a few Li ions (of type 1, charge +1) to balance the charge? Or is there a more proper way?

Thank you very much for your advice！

2

Very serious. You have created a system that would in real life explode instantly.

You cannot ignore this warning unless the magnitude of the net charge is below 1.0 and thus the net charge the result of rounding and not due to construction of the system.

You have to adjust the settings for creating the geometry so that it is charge neutral by construction.

1 Like