 Re: Dipole correction for slab calculation ( No.1 ) |
- Date: 2014/08/11 19:17
- Name: T. Ozaki
- Hi,
The dipole-dipole interaction between slabs can be eliminated by
the effective screening medium (ESM) method, whose usage can be
found at
http://www.openmx-square.org/openmx_man3.7/node118.html
An application of the ESM method to slab has been demonstrated in
http://arxiv.org/abs/1407.2698
Regards,
TO
|
| Re: Dipole correction for slab calculation ( No.2 ) |
- Date: 2014/08/13 18:05
- Name: Sung Beom Cho <csb444@hanyang.ac.kr>
- Thank you for your kind reply.
I just tested the ESM method and found it works fine. It means that the electric field option is completely under zero-field external electric field. From the written Hartree potential, I checked that the dipole-dipole interaction between the slabs is clearly eliminated. Furthermore, the electric field with "on4" boundary condition works fine.
I also checked if scf.Electric.Field option works with the ESM method. From the calculated hartree potential, I found out that the scf.Electric.Field option does not work with the ESM method for all kind of boundary conditions. Thus, the scf.Electric.Field cannot describe the correct electric field effect and the ESM method is required.
However, because the ESM method limits the a-axis, it is hard to apply to NEGF calculation. To simulate FET transistor with NEGF method, bias voltage should be applied in a-axis and gate voltage should be applied in c-axis. Thus, I think scf.Electric.Field to c-axis with dipole correction is necessary for NEGF calculation. Is there anyway to avoid the dipole-dipole interaction with scf.Electric.Field option?
Best Regards,
Sung Beom Cho
 |
| Re: Dipole correction for slab calculation ( No.3 ) |
- Date: 2014/08/19 21:40
- Name: T. Ozaki
- Hi,
The ESM method is not compatible with the NEGF method.
Although Ver. 3.7 does not have functionality to correct the dipole-dipole interaction
when electric field is applied along c-axis in the NEGF method, it may be theoretically
possible to combine two boundary conditions: the boundary condition of Hartree potential
in the NEGF method along a-axis and the boundary condition by the ESM method along c-axis.
As long as I know, there might be no such implementation, and it may be worth considering.
Regards,
TO
|
| Re: Dipole correction for slab calculation ( No.4 ) |
- Date: 2014/08/20 11:08
- Name: Sung Beom Cho <csb444@hanyang.ac.kr>
- Thank you, Ozaki.
Combination of ESM and NEGF would be great and worth for lots of applications.
However, I think simple dipole correction might be enough for vacuum-containing supercell. This theory is well established in [PRB 59 12301]. Lots of plane-wave codes like VASP already have this functionality to calculate the polarized slab.
I think this approach might be much easier than the combination of the ESM. In most NEGF calculation, I think it would be enough to satisfy the boundary condition in the neutral slab. Because the calculation cost of the vacuum is almost free in localized basis set, I think this method is very useful for lots of systems.
I wonder if this functionality is already implemented in openMX, or in plan to be implemented.
Best Regards,
Sung Beom Cho
|
| Re: Dipole correction for slab calculation ( No.5 ) |
- Date: 2014/08/21 16:51
- Name: T. Ozaki
- Hi,
> I wonder if this functionality is already implemented in openMX,
The functionality has never been implemented in OpenMX.
> or in plan to be implemented.
It is worth considering.
Regards,
TO
|
This thread is locked.Only browsing is available.
Dipole correction for slab calculation