Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Concerns issues with computing quasiparticle corrections to the DFT eigenvalues - i.e., the self-energy within the GW approximation (-g n), or considering the Hartree-Fock exchange only (-x)

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

Post Reply
Bikash
Posts: 7
Joined: Mon May 17, 2021 1:28 pm

Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Post by Bikash » Thu Apr 21, 2022 2:02 pm

Dear developers,

I am running the exchange self-energy part for monolayer WTe2 following the instruction
given here - http://www.yambo-code.org/wiki/index.ph ... rtree_Fock.
Although I did not change the value of VXCRLvcs, still I am getting different
value of E_xc printed in the QE SCF output file and Yambo report file.
In QE scf file - E_xc= -261.64702929 Ry, whereas in Yambo report file- E_xc=-109.3980 Ry.
I checked in the example hBN-2D file these two values are the same.
Could you please point out my mistake?
You do not have the required permissions to view the files attached to this post.
Bikash Patra
TIFR, Mumbai, India.

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Post by Daniele Varsano » Thu Apr 21, 2022 2:22 pm

Dear Bikash,

this usually happens when you use pseudopotentials with non-linear core correction (e.g. pseudodojo).
If this is the case you need to activate them in Yambo using UseNLCC keyword.
http://www.yambo-code.org/wiki/index.ph ... es#UseNLCC

In this way, the nlcc are correctly included in the <Vxc>. The problem is then that there is no way to include them in the HF part and it is not clear
what's the better procedure, ie not including them at all or including them in the Vxc only. This is the same reason why these pseudopotentials are not
suggested in QE when using hybrid functional. You can try to conclude them, or think about using PSP without nlcc e.g. sg15 (http://www.quantum-simulation.org/potentials/sg15_oncv/)

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Bikash
Posts: 7
Joined: Mon May 17, 2021 1:28 pm

Re: Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Post by Bikash » Fri Apr 22, 2022 6:31 am

Dear Sir,

Thank you for your suggestions.
Now if I use the sg15 PSP without nlcc, the Exc value in
the QE scf output file and Yambo report file match for nosoc calculation.
If I include SOC, these two values again differ.
In QE scf file - E_xc= -159.31593215 Ry, whereas in Yambo report file- E_xc= -109.3966 Ry.
You do not have the required permissions to view the files attached to this post.
Bikash Patra
TIFR, Mumbai, India.

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Post by Daniele Varsano » Fri Apr 22, 2022 7:40 am

Dear Bikash,

the psuedopotential you are using seems to have nlcc

Code: Select all

xc Functional : Perdew, Burke & Ernzerhof(X)+Perdew, Burke & Ernzerhof(C)
 [xc] LIBXC used to calculate xc functional
  PP Number of Atomic Species                      :  2
  NLCC RL compoments                               :  115637
In order to match the QE value, you need to include them. You can check by opening the PSP file and check if the label core_correction is true or false.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Bikash
Posts: 7
Joined: Mon May 17, 2021 1:28 pm

Re: Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Post by Bikash » Fri Apr 22, 2022 7:54 am

Dear Sir,
Thank you for your answer.
In both the pseudopotential the core_correction tag is false.
Another question, even if they do not match, can I proceed with
the calculation?

with regards,
Bikash.
You do not have the required permissions to view the files attached to this post.
Bikash Patra
TIFR, Mumbai, India.

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Problem in understanding the Exc value printed in HF calcution and DFT calculation.

Post by Daniele Varsano » Fri Apr 22, 2022 11:59 am

Dear Bikash,

be sure you are using the correct SAVE directory as the report you posted indicated that nlcc were present and you were obtaining a value similar to the previous calculation obtained with nlcc PSP.
Be sure also to remove the ndb.HF* database before repeating the calculation.
Another question, even if they do not match, can I proceed with the calculation?
Well, if the core correction is not present in the PSP, the QE and Yambo value of E_xc should be the same.
Note that E_xc is not used by Yambo, it is written as a check and if they do not agree it is possible that could have a problem with the <Vxc> term.
As I wrote before you can also use PSP with NLCC but then it is not clear if it is better to include them in Yambo or not.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Post Reply