negtive HF corrections when including SOC

Deals with issues related to computation of optical spectra, solving the Bethe-Salpeter equation.

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano

Post Reply
arlonne
Posts: 26
Joined: Thu Apr 18, 2013 4:50 pm

negtive HF corrections when including SOC

Post by arlonne » Tue May 13, 2014 4:19 pm

Dear all,

I found a strange thing when I include spin-orbit coupling (SOC) in G0W0 calculations. The HF correction without SOC opens a large band gap comparing with GGA-PBE, however after including SOC, the HF gives a negative band gap as follows shown.

without SOC @ Gamma k-point:
<159|HF|159> = -11.62607 0.1017E-8 <159|DFT|159> = -10.84202 0.000000
<160|HF|160> = -10.28354 -.3978E-8 <160|DFT|160> = -10.42272 0.000000 HOMO
<161|HF|161> = -5.809332 0.262E-09 <161|DFT|161> = -8.814879 0.000000 LUMO
<162|HF|162> = -5.856881 -.242E-09 <162|DFT|162> = -8.845275 0.000000

with SOC @ Gamma:
<319|HF|319> = -7.648899 -.4403E-8 <319|DFT|319> = -9.515832 -.1267E-7
<320|HF|320> = -7.648894 -.1449E-8 <320|DFT|320> = -9.515823 0.000000 HOMO
<321|HF|321> = -9.057126 0.540E-09 <321|DFT|321> = -10.81462 0.6336E-8 LUMO
<322|HF|322> = -9.057129 0.2195E-8 <322|DFT|322> = -10.81461 0.000000
And the final G0W0 with SOC are:
B=319 Eo= -0.08 E= 1.32 E-Eo= 1.41 Z=0.81 nlXC=-7.648899 lXC=-9.515832 So=-0.129964
B=320 Eo= -0.08 E= 1.32 E-Eo= 1.41 Z=0.81 nlXC=-7.648894 lXC=-9.515823 So=-0.132115
B=321 Eo= 0.08 E= 1.56 E-Eo= 1.48 Z=0.82 nlXC= -9.05713 lXC=-10.81462 So= 0.04808
B=322 Eo= 0.08 E= 1.56 E-Eo= 1.48 Z=0.82 nlXC= -9.05713 lXC=-10.81461 So= 0.05031

The G0W0 gives a similar band gap to PBE because of negative HF corrections? I also checked other k-points, the HF opens positive gaps between band 320 and 321. Considering some parameters are important to G0W0, I have tried several calculations with different GbndRnge and NGsBlkXp, however, G0W0 still around 0.2 eV, which is significantly smaller than the experimental value of about 1.2 eV. So, Is there something wrong with my calculations or Yambo can not perform G0W0 with SOC?

P.s. This is a perovskitic structure (ASnX3). The PBE band gap is 0.28 eV; PBE-SOC band gap is 0.16 eV. The pwscf generated band gap is compatible with other DFT codes.
The attachments are the main output of calculations with/without SOC by YAMBO.
You do not have the required permissions to view the files attached to this post.
arlonne
Department of Physics and Astronomy
Northwestern University, IL 60208

User avatar
Davide Sangalli
Posts: 610
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy
Contact:

Re: negtive HF corrections when including SOC

Post by Davide Sangalli » Tue May 13, 2014 5:03 pm

Dear Arlonne.
It is hard to say from just the report files.

The only comment I have is that the number of RL vectors seems to be different between the two runs. Is there a reason ?

Can you reproduce this result with a smaller calculation?
Try for example with the isolated atoms or the bulk version of the atms which compose your system: Sn,I,C,N,H

If you find the same problem in a small system I'll try to exploit it

Best regards,
Davide
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
https://sites.google.com/view/davidesangalli
http://www.max-centre.eu/

arlonne
Posts: 26
Joined: Thu Apr 18, 2013 4:50 pm

Re: negtive HF corrections when including SOC

Post by arlonne » Tue May 13, 2014 5:33 pm

Thank you for your attention. I have tried different RL vectors, however the result is not dependent on the number of vectors, I mean 5000-10000 all generate same results.
So, SOC is implemented into GW and BSE? If it is, I will try a small case.

Thank you very much!
arlonne
arlonne
Department of Physics and Astronomy
Northwestern University, IL 60208

User avatar
Davide Sangalli
Posts: 610
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy
Contact:

Re: negtive HF corrections when including SOC

Post by Davide Sangalli » Tue May 13, 2014 7:39 pm

Yes, both BSE and GW (within the G0W0 approximation) are implemented with the full spinorial dependence of the wave-functions.
This way the SOC coupling is included in both the calculations.

If you are interested you may want ot have a look to this reference as an example of a BSE calculation.
http://journals.aps.org/prb/abstract/10 ... .88.045412

Best regards,
Davide
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
https://sites.google.com/view/davidesangalli
http://www.max-centre.eu/

Post Reply