Issue about local electric field strength calculation

[xiaocheng_zeng2]

Hello everyone,

I am currently working on calculating the local electric field of a system, as shown in the attached figure.

Based on Bader charge analysis and charge density difference calculations, Region I has gained electrons and carries a negative charge, while Region II has lost electrons and carries a positive charge.

cdc.png

Following the discussion in this post: https://ww.vasp.at/forum/viewtopic.php? ... 3e3ff95b40, I learned that the electrostatic potential recorded in VASP’s LOCPOT file is referenced to electrons.

Therefore, I used the formula E=−Δ(LOCPOT/−e)/Δz to calculate the electric field strength along the z-direction.

locpot.png

The results show that the average electric field strength in the yellow region between Region I and Region II is positive, meaning the electric field points along the positive z-direction.

However, this is opposite to the direction predicted by the parallel plate capacitor model, where the electric field is expected to point from the positive to the negative charge.

plate.png

I would like to ask:
1. Is there any issue with my method for calculating the electric field strength?

2. What might be the reason for the discrepancy between my analysis and the parallel plate capacitor model? Is the parallel plate capacitor model applicable in this case?

Thank you very much!

[fabien_tran1]

Hi,

I have moved your questions to "From users to users" since they may be more related to physics rather than VASP usage. Other users may have experience with this topic. Concerning LOCPOT, I suppose that it was generated with LVHAR=.TRUE. such that it contains only the electrostatic component, right?

[xiaocheng_zeng2]

Hi,

I have moved your questions to "From users to users" since they may be more related to physics rather than VASP usage. Other users may have experience with this topic. Concerning LOCPOT, I suppose that it was generated with LVHAR=.TRUE. such that it contains only the electrostatic component, right?

Yes, the LVHAR = .TRUE. setting has been applied and VASP version is 6.3.0. I have also uploaded the input file. Did I do something wrong?

[xiaocheng_zeng2]

vasp-test.zip

[fabien_tran1]

Hi,

At first sight I did not see anything wrong in your INCAR file, although it is difficult to provide specific recommendations due to my lack of expertise in this topic.

Actually, if I interpret correctly your plot "electric_field_fixed", it seems that there is also a non-negligible negative portion in the yellow box. What I am wondering is how sensitive is the calculated electric field with respect to the various choices for the calculation, in particular the exchange-correlation functional and POTCAR.