Electric field and HOMO-LUMO gap questions....help

[bendix]

I recently met 2 questions following these....

1st: Electric field question
1.EF unit is V/A or eV/A?
OUTCAR output eV/A, but I think EF unit is V/A.

2.Can EF calculate in SC?
I calculate EF in SC, but electron steps don't convergence over 60 steps. So PRB vol.64 pp.125403 called this behavior "charge sloshing". It suggest to calculate EF in non-SC. Really???

3.The same, PRB vol.64 pp.125403 refer to set negative EF value (IDIPOL=-1) not positive in VASP, this means electrons will tunnel into vacuum from slab. Really???


2nd: HOMO-LUMO gap questions
Whether PROCAR can calculate HOMO-LUMO?
I try to type "grep band PROCAR", can look under~
band 120 # energy -5.56448401 # occ. 2.00000000
band 121 # energy -3.91086958 # occ. 0.00000000

HOMO-LUMO gap calculte difference between band120 and band121 or not.

If not, how to obtain HOMO-LUMO gap?

Thanks~~
<span class='smallblacktext'>[ Edited ]</span>

[bendix]

please help me~~
I wait so long.

[admin]

1) the (+ or -) direction of the field has to be given via EFIELD, not by setting IDIPOL to add a field in -x direction, please
give
IDIPOL = 1; EFIELD = -xx (the field strength)
if IDIPOL+x), EFIELD is not read at all from INCAR.
To check, please have a look which value your field was assigned (grep for EFIELD in OUTCAR, you will probably not find any entry).

2) as DFT is a ground state theory, the HOMO-LUMO gap is not defined correctly by taking the energy differences of the Kohn-Sham levels because the KS-energy of the LUMO will not be correct (note that usually DFT severely underestimates gap-widths).

[bendix]

thanks for your result!

May admin also answer the second important question?

Can EF calculate in SC?
I calculate EF in SC, but electron steps don't convergence in 60 steps. So PRB vol.64 pp.125403 called this behavior "charge sloshing". It suggest to calculate EF in non-SC. Really???

[ccmin4]

So, the EF unit is V/A or eV/A?