GW QP Bandgap

[Daniele Varsano]

Dear Reza,

the GW calculation seems to be reasonable. You can have a look at your KS and QP band structure by plotting the O_job_qp.txt file:
e.g

Code: Select all

>gnuplot
p "O_job_qp.txt" u 1:3 w p t "KS"
rep "O_job_qp.txt" u 1:($3+$4) w p t "QP"

Note that the zero is set to KS VBM, so you would need to shift the QP one.

I do not understand anyway what went wrong with the ypp interpolation.

Best,
Daniele

[Reza_Reza]

Thanks,
Does it mean if I simply plot the GW band structure in gnuplot using
p "O_job_qp.txt" u 1:($3+$4) w p t "QP"
without ypp calculation??

How should I find the value i shift that should be applied in the QP band structure?

Thank You and Best,

[Daniele Varsano]

Dear Reza,

Does it mean if I simply plot the GW band structure in gnuplot using
p "O_job_qp.txt" u 1:($3+$4) w p t "QP"
without ypp calculation??

Yes, you have already calcualted QP along the path so interpolation is not needed.

How should I find the value i shift that should be applied in the QP band structure?

It is totally arbitrary, it is a semiconductor, you can set the Fermi Energy at the valence band maximum (as done by Yambo for the KS bands), or in the middle of the gap.

Best,
Daniele

[Reza_Reza]

Hello,

I am attempting to converge the EXXRLvcs. However, the band gap computed at the HF and DFT levels remains the same, approximately 1.24 eV. Below is my input file:

HF_and_locXC
EXXRLvcs=$value
VXCRLvcs= 900000
%QPkrange
41|41|66|67|

In the report file, the direct band gap is at k-point 41 (attached). However, for HF, the direct band gap is at k-point 40. When I adjust the k-points to capture 40 (40|41|66|67|) without any other changes in the input file, the HF direct band gap shifts to k-point 39. Similarly, when I adjust it to 39, it becomes 38. Despite these adjustments, the band gap value in the report file remains consistent.

Attached are the report_HF.txt and output.hf.txt files for your review.

Best regards,

Best,

[Daniele Varsano]

Dear Reza,

as I told you in a previous post, if you calculate the correction to a limited number of k points, you need to look at the calculated correction and *NOT* the gap as reported in the report. This is because if you do not calculate the correction in all the BZ it cannot be correctly calculated as you have the energies corrected for some kpoints, and KS energy for others. This is the reason why the k point where the gap is located it shifts. You can anyway calculate the gap at the given k point looking at the output file.

Hope now it is more clear,

Best,
Daniele