[Wannier] Asymmetric value of AHC

[박민규]

Dear Stepan,

Thank you for the comment and suggestion.
I’ll have a look.

Best,
============================================
Minkyu Park
Research Institute of Basic Sciences, University of Ulsan,
93, Daehak-ro, Nam-gu, Ulsan, 44610 Republic of Korea
minkyupark at ulsan.ac.kr<mailto:minkyupark at ulsan.ac.kr>
+82-52-259-1473
============================================

On 8 Jun 2020, at 1:17 AM, Stepan Tsirkin <stepan.tsirkin at uzh.ch<mailto:stepan.tsirkin at uzh.ch>> wrote:

Dear Minkyu,

In general, the MLWFs often happen to slightly break the symmetry. (Unless symmetry-adapted WFs are used. However SAWFs are not implemented with disentanglement so far) So, for the example of Fe, it is normal that a small x and y component appears.

Regarding v3.0, I am not sure what’s happening. There were some changes between versions, to improve performance, but those should not affect the physical results. Probably for a denser grid it will give a similar result as the v3.1 .  50x50x50 is actually a coarse grid for ahc, and ahc may strongly depend on the fermi level (use fermi scan), especially for coarse grids the calculation yields huge unphysical peaks.

Symmetries are explicitly accounted in the wannier-berri package (see on github):  It also has more improvements over postw90.x , which make the calculations of AHC (and other quantities) much faster and more accurate. I advise you to try it.

Best regards,
Stepan.

------------------------------------
Stepan Tsirkin
University of Zurich
stepan.tsirkin at uzh.ch<mailto:stepan.tsirkin at uzh.ch>
<wannierberri.png>
https://github.com/stepan-tsirkin/wannier-berri

On 7 Jun 2020, at 17:52, 박민규 <minkyupark at ulsan.ac.kr<mailto:minkyupark at ulsan.ac.kr>> wrote:

Dear all,

While I tried to follow example18 of Wannier90 v3.1 with QuantumEspresso v6.5, I found asymmetric value between AHC_x and AHC_y.
The result is following.

   * Anomalous Hall conductivity

 Interpolation grid: 25 25 25

 Fermi energy (ev):   12.6438

 AHC (S/cm)       x          y          z
 ==========    -0.0003    -0.0063   552.3105

Or if I increase grid,

   * Anomalous Hall conductivity

 Interpolation grid: 50 50 50

 Fermi energy (ev):   12.6438

 AHC (S/cm)       x          y          z
 ==========    -2.3241     0.0155   648.0428

According to the solution_booklet, AHC_x and AHC_y in this case should be zero unless we turn on adaptive refinement grid. If we turn it on, they should agree at least its absolute value.
As you see, my result does not have such restriction on AHC_x and AHC_y.

Is this because of numerical error? or something has changed since the example was written?

Best regards,

P.S. If I do the same calculation with Wannier90 v3.0 then the result looks like this.

   * Anomalous Hall conductivity

 Interpolation grid: 25 25 25

 Fermi energy (ev):   12.6438

 AHC (S/cm)       x          y          z
 ==========    -0.9038    -0.9338   497.5371


   * Anomalous Hall conductivity

 Interpolation grid: 50 50 50

 Fermi energy (ev):   12.6438

 AHC (S/cm)       x          y          z
 ==========   -45.7805   -44.3405  1049.8894

So the value itself changes depending on which version I use.

============================================
Minkyu Park
Research Institute of Basic Sciences, University of Ulsan,
93, Daehak-ro, Nam-gu, Ulsan, 44610 Republic of Korea
minkyupark at ulsan.ac.kr<mailto:minkyupark at ulsan.ac.kr>
+82-52-259-1473
============================================
_______________________________________________
Wannier mailing list
Wannier at lists.quantum-espresso.org<mailto:Wannier at lists.quantum-espresso.org>
https://lists.quantum-espresso.org/mailman/listinfo/wannier


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20200607/5a6f8e40/attachment-0001.html>