LDIPOL activation for charged slab

[kazem_zhour]

Dear support,

I am studying the adsorption of molecule on charged slab (of course LDIPOL need to be activated in this case). When I consider the neutral slab, VASP finish the job without any error, but when I move to my charged slab by just decreasing NELECT and not changing anything else in the INCAR, the calculations stopped from the beginning with an error announcing that "LDIPOL = .TRUE. must be selected only for cubic supercell". Even though I am using for the charged slab the same POSCAR I used for the neutral slab!
It would be great if you can provide me with some advices that can help me to solve this problem. Thanks for considering this issue, and I’ll wait to hear from you.

Sincerely

Kazem

[andreas.singraber]

Dear Kazem,

welcome to the VASP forum! Could you please attach your input and output files according to the forum posting guidelines.

Thank you!

Best,
Andreas Singraber

[kazem_zhour]

Dear Andreas,

Thank you for your replay. Kindly find attached the needed files.

Best,
Kazem

[kazem_zhour]

I am still interested in knowing the answer of this issue please.

[andreas.singraber]

Dear Kazem,

I'm sorry that your question was not answered yet, I will have a look at it...

Best,
Andreas

[andreas.singraber]

Dear Kazem,

I had a look at your problem and it seems that LDIPOL = .TRUE. for charged systems is only implemented for cubic supercells. There is also a note on the Wiki page saying:

Restrictions: For charged systems, the potential correction is currently only implemented for cubic supercells. VASP will stop if the supercell is not cubic and LDIPOL=.TRUE.

Maybe you could try to make the box cubic.. currently the lattice vector pointing out of the slab has a length of 30 Angstrom. The in-plane lattice vectors have a length of about 17.8 A. If you decrease the vacuum thickness a bit (so that you still do not have an effect from periodicity) while at the same time extending the slab in the other two directions you may be able to get a cubic box. Then, the LDIPOL feature should work... I hope that is a feasible option for you?

All the best,
Andreas Singraber