 Re: BerryFlux in slandered output file of calB ( No.1 ) |
- Date: 2020/01/17 18:40
- Name: Hikaru Sawahata <sawahata@cphys.s.kanazawa-u.ac.jp>
- Dear Umar Farooq
Thank you for the question.
BerryFlux \Phi means
\Phi = \oint {\bf A} \cdot d{\bf k}
Here, {\bf A} is the Berry connection. Berry Flux is the dimensionless quantity.
In the Dirac cones, Berry Flux becomes \pm \pi.
In the Fukui-Hatsugai-Suzuki method, the values of F({\bf k}) = Im log UUUU become "BerryFlux", not "Berry curvature", so this code divide these values by "dk^2" and outputs to "BerryC*.dat".
(http://openmx-square.org/openmx_man3.9/node184.html)
For confirming the calculation working correctly or not, this code outputs
<Berry Flux> "k1", "k2", "BerryFlux"
to standard output, here k1, k2 indicate k = k1G1 + k2G2 ( G_n is a reciprocal lattice vector).
If you want to obtain the "Berry curvature", please see the output file "BerryC*.dat".
I'm sorry to confuse you.
Kanazawa University
Hikaru Sawahata
|
| Re: BerryFlux in slandered output file of calB ( No.2 ) |
- Date: 2020/01/20 13:16
- Name: Umar farooq <farooqmu@sustech.edu.cn>
- dear,
First of all thank for your reply that clears up a lot for me. I have two more questions regarding these calculations.
1. l have semiconducting system with 400 electron when I calculate I got an out put file named as BerryC400.dat with input file calB.in
1
0
0 0 1
50 50
However, I wanted to do some charge doping, in case of small hole doping of ie. g0.2h I got BerryC399.dat. In the DFT SCF calculations *.out file I have Number of States = 399.80000000000007
does that mean the 400th band which is partially occupied is not included in the in berry curvature and ChernNumber calculation if so what is your suggestion? How can I mimic this small charge doping.? Is it fine to handle metallic systems in berry curvature calculations?
Maybe I am bit naive.
One way is that I can just simply control number of bands in the berry curvature calculations to simulate charge. However, in my system the electronic structure and magnetic moment is very responsive to the charge doping.
2. It is about the format of input file according to manual:
1st line g1h mean sum of the Berry curvature of all the bands O mean Berry curvature is calculated for a specific band.
but the 2nd line according to the manual, "In the second line, you need to specify how many bands are taken into account for the calculations, .... "
If I mention 0 in first line and 400 in 2nd line does it mean I am only calculating for 400th atom ?
M. Umar Farooq
Southern University of Science and Technology
Department of Physics
1088 Xueyuan Avenue, Nanshan District, Shenzhen, Guangdong
 |
| Re: BerryFlux in slandered output file of calB ( No.3 ) |
- Date: 2020/01/22 02:56
- Name: Hikaru Sawahata <sawahata@cphys.s.kanazawa-u.ac.jp>
- Dear,
1. In your case, I think you have to calculate Berry curvatures in metallic state, in other words, you have to calculate Berry curvatures at Fermi energy. Unfortunately, this code gcalB.ch cannot compute it. Ifm developing and tuning the code which can compute Berry curvature and anomalous Hall conductivity of metallic system now.
If you are in a hurry, my suggestion is using gWannier90h code. It can compute Berry curvature every chemical potential. Please refer this manual of interface with Wannier90.
http://www.openmx-square.org/openmx_man3.9/node148.html
2. Your understanding is correct, my explanation is incorrect, sorry.
>If I mention 0 in first line and 400 in 2nd line does it mean I am only calculating for 400th atom ?
Kanazawa University
Hikaru Sawahata
|
| Re: BerryFlux in slandered output file of calB ( No.4 ) |
- Date: 2020/01/23 16:24
- Name: Umar farooq <farooqmu@sustech.edu.cn>
- Dera Hikaru Sawahata,
Thank you very much,
I am trying Wannier90. Actually to converge Wannier functions on the DFT data is quite laborious, in some cases. Technically we should be able to get hall conductivity berry curvature and related properties with much more easily Directly from DFT data, especially with the localized base set codes (OpenMX).
I eagerly want to try new code for berry curvature even if it is in initial stage.
M. Umar Farooq
Southern University of Science and Technology
Department of Physics
1088 Xueyuan Avenue, Nanshan District, Shenzhen, Guangdong
|
| Re: BerryFlux in slandered output file of calB ( No.5 ) |
- Date: 2020/02/17 23:29
- Name: Arqum Hashmi <hashmi.arqum.ga@u.tsukuba.ac.jp>
- Dear Hikaru Sawahata,
I have some questions regarding berry curvatures
1st: I want to calculate the berry curvature of a single specific band (whether system is semiconductor or metallic) then is it ok to calculate it by calB.c?? Technically can calB.c code calculate it correctly for a specific band.
2nd : the second line of calB.in means "how many bands are taken into account for the calculations" if i have 100 bands and i select 1 in first line and 10 in 2nd line then it means it will sum the from 1-10 bands. Am i right??
3rd: In manual you said, Berry curvature can be calculated for set of disentangled bands, for example i have 100 bands and i want to calculate the berry curvature for just 50-55 then can we do it?? if we can then how can we mention it in calB.in??
Thank you in advance.
|
| Re: BerryFlux in slandered output file of calB ( No.6 ) |
- Date: 2020/03/28 20:13
- Name: Hikaru Sawahata <sawahata@cphys.s.kanazawa-u.ac.jp>
- Dear Arqum Hashmi
Thank you for question,
1. calB.c can compute the berry curvature of a single specific band, but it must be isolated from another bands. This code can calculate only "a disentangled bands".
2. Your understanding is correct.
3. That option is not implemented in this version, but it is easy to do by modifying the part of determinant calculation in the code.
Hikaru Sawahata
|
|
This thread is locked.Only browsing is available.
BerryFlux in slandered output file of calB