QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: NW on April 7, 2020, 14:41
-
Hello,
I calculated the band structure of bilayer graphene and the results with the script file were attached here. When I zoomed on the bands near to Fermi energy, it is clear that the Fermi level is not somewhere between the valance and conduction bands. How can I set the Fermi energy in the gap of the bilayer?
Thanks
-
This tiny shift looks very much within numerical noise. You cannot just set the Fermi level, as the Fermi level position is actually computed, and this is where this noise comes from. You may try increasing the k-point sampling to increase the accuracy.
-
I agree with Petr, this is within the "noise level", but of course you can control what that level is by increasing the k-point sampling (making sure to keep the K point included) and lowering the electron temperature (and possibly converge the SCF loop a bit harder). You could even get very creative possibly and add extra k-points around the K point instead of just doing a blanket Monkhorst-Pack sampling around the whole Brillouin zone, but then you also need to ask why it's important to have Fermi level so precisely determined (other sources of error will be larger).
Btw you mentioned "it is clear that the Fermi level is not somewhere between the valance and conduction bands", well, that would be hard since the system appears to be metallic.
-
Dear Anders,
Thank you so much for your specific solution.
My question is how can I add extra k-points around K?
-
Alas, I spoke too soon. This is only a feature planned for future release... You can do it for transmission spectra for transport calculations, but not yet for the self-consistent loop.