G0W0+PPA Terminator Behavior

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
AzulineSky
Posts: 9
Joined: Tue Oct 18, 2022 10:25 pm

G0W0+PPA Terminator Behavior

Post by AzulineSky » Tue Apr 25, 2023 3:00 pm

Dear all,

I'm currently performing some convergence studies for G0W0-PPA calculations for Si and LiF and noticed some weird behavior when setting XTermKind ="BG" and GTermKind = "BG". The attached plots show the convergence studies of the direct bandgap of Si and LiF for the parameter BndsRnXp, GbndRnge and NGsBlkXp. As shown on the wiki I converged NGsBlkXp and BndsRnXp simultaneously (W_conv) and GbndRnge in separate convergence study (G_conv). I'm wondering why the resulting direct bandgaps with and without the terminators differ so much in both cases and why the direct bandgaps don't seem to converge to the same value. Also the direct bandgap of LiF with XTermKind ="BG" and GTermKind = "BG" is much smaller than expected and round 1eV smaller than the direct bandgap calculated with the same settings and XTermKind ="none" and GTermKind = "none".

Best regards,
Max
You do not have the required permissions to view the files attached to this post.
Max Großmann
TU Ilmenau
Germany

AzulineSky
Posts: 9
Joined: Tue Oct 18, 2022 10:25 pm

Re: G0W0+PPA Terminator Behavior

Post by AzulineSky » Tue Apr 25, 2023 3:01 pm

Missing Plot
You do not have the required permissions to view the files attached to this post.
Max Großmann
TU Ilmenau
Germany

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

Re: G0W0+PPA Terminator Behavior

Post by Daniele Varsano » Fri Apr 28, 2023 8:52 am

Dear Max,

we should have a close look at this to identify what's going wrong here. Probably some missing convergence in NGsBlkXp or maybe some non optimal position of the high energy pole.
My suggestion, in general, is to avoid to use XTermKind, as it is not always stable and memory intensive.
I suggest you to repeat your test using GTermKind only.

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