Calculating migration energy barrier in presence of eternal electric field in charged system

[tmp_BPNUgJCrKRoP4vStxq6QzPlo]

I'd like some help understanding a few issues I've faced. I want to calculate the migration energy barrier for a metal ion (e.g., Cu⁺) in an oxide (e.g., SiO₂) system. Specifically, I want to study how an external electric field affects the migration behavior of metal ions in oxides. To do this, I ran NEB simulations for both a neutral system (no added/removed charge) and a charged system (one fewer electron to make Cu⁺), each in the presence of an external electric field (e.g., 0.1 eV/Å).
For the neutral system, I used the following INCAR tags:
LDIPOL = .TRUE.
IDIPOL = 3
DIPOL = 0.5 0.5 0.5
EFIELD = 0.1
The simulation run and it calculates the migration behavior.
However, for charged system I use this INCAR:
NELECT = (total number of electrons) - 1
LDIPOL = .TRUE.
IDIPOL = 3
DIPOL = 0.5 0.5 0.5
EFIELD = 0.1
when I include the charge into the supercell system, the simulation did not progress and stops immediately, saying "Warning: For charged systems, the potential correction is currently only implemented for cubic supercells. VASP will stop if the supercell is not cubic and LDIPOL is used." Then, I turned off the LDIPOL and run the simulation. The INCAR tags I used was:
NELECT = (total number of electrons) - 1
LDIPOL = .FALSE.
#IDIPOL = 3
#DIPOL = 0.5 0.5 0.5
EFIELD = 0.1
Now I have few questions regarding the methodology to calculate the migration behavior of ions in presence of external electric field.
1. When we calculate a migration barrier, are we calculating the migration barrier of an ion, or simply of an atom (with the electronic structure relaxing self-consistently around it)?
2. Does removing (or adding) one electron from the supercell correspond to creating an ion at the site I intend, or does it just add charge to the system as a whole? What is the physical effect of adding this charge? I now doubt that adding a charge to the supercell system does not have anything to do with making Cu+ ions. Can you please help me clarify it?
I have included the INCAR files for both neutral and charged system here.

[fabien_tran1]

Hi,

Indeed, LDIPOL is implemented only for cubic cells. However, note that for systems with reduced dimensionality a possible alternative is to use the truncation method (see also https://vasp.at/wiki/Category:Electrostatics), which is available since VASP.6.5.0.

When an electron is added or removed, no constraint is applied, i.e., the charge can move and distribute freely. Where the charge goes is determined by the physics (the variational principle).

By setting LORBIT=11, which you did, the charges on each atom is printed in OUTCAR (search for "total charge"). Then, you compare the charge on the Cu atom between the two calculations.