self consistency on wave functions

Various technical topics such as parallelism and efficiency, netCDF problems, the Yambo code structure itself, are posted here.

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano, Conor Hogan, Nicola Spallanzani

Post Reply
manaswitakar
Posts: 21
Joined: Wed Nov 23, 2022 12:20 pm

self consistency on wave functions

Post by manaswitakar » Tue Aug 29, 2023 4:03 pm

Hi,

I am currently trying to perform full self consistent GW calculations on double perovskites. The calculations are converging, however, the final GW band gaps are still dependent on the xc functional used. For example; using PBE, I get the final (after self consistency convergence) band gap of 2.65 eV, whereas with B3LYP, the band gap is 3.44 eV. I think this is because the self-consistency is only on Eigen values and not on the wave functions. Is there a way I can perform self consistency on wave functions using YAMBO? Are these calculations too expensive for perovskites? Please let me know.

Thanks and Best Regards,
Manaswita
Manaswita Kar

Sorbonne University, CNRS,
Paris, France

User avatar
Daniele Varsano
Posts: 3980
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: self consistency on wave functions

Post by Daniele Varsano » Tue Aug 29, 2023 4:20 pm

Dear Manaswita,

are you doing eigenvalue self-consistency on both G and W, or G only (GW0)?
My suggestion is to go for GW0, but of course there is not a rule for that.
At the moment, self-consistency on wfs can be done only at COHSEX level of theory and yes, these calculations are rather expensive.

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/

manaswitakar
Posts: 21
Joined: Wed Nov 23, 2022 12:20 pm

Re: self consistency on wave functions

Post by manaswitakar » Tue Aug 29, 2023 4:36 pm

Hi Daniele,

Thanks for your response. Currently, I am doing GW (self-consistency on both G and W). I will try GW0 and check how much difference it makes for the perovskites.

Best Regards,
Manaswita
Manaswita Kar

Sorbonne University, CNRS,
Paris, France

Post Reply