Highest occupied level is not correctly obtained from r_setup

[Bramhachari Khamari]

Dear Developers,
I would like to calculate GW band for a material. Ground state calculation (QE) gives the highest occupied level as 7.8793 eV (Occupation=fixed is used). When, I perform yambo to get the r_setup, I obtain the following information

Fermi level=7.0597 eV
conduction band minimum=7.0597 eV
Valence band maximum=7.0597 eV
I have following two questions in my mind

i)Why I am not getting the highest occupied energy same as my ground state value obtained using QE. Any help will be appreciated.
ii) Why I see metallic bands (number 17, 18) in my r_setup although, I am not getting fractional occupation at any k-point (scf.out from QE).

Attached is the r_setup file.



Regards,

[Daniele Varsano]

Dear Bramhachari,

in a semiconductor yambo sets the VBM at 0eV, so you will not have the same VBM and CBM of QE.
Anyway, here the problem is that your system is seen as metallic.
Please check, and eventually post the QE output to see if there is something odd, e.g. unoccupied bands with lower energies than occupied ones (sometimes this can happen when forcing fixed occupation in metallic systems.)
In general, there should be a one to one correspondence between QE and Yambo energies beside a rigid shift (check if this is the case or not!).

Best,
Daniele

[Bramhachari Khamari]

Dear Sir,
Thanks for the reply. From the out put I see that unoccupied energy(6.6773) is lower than highest occupied one(7.8793) as you mentioned. In literature I found that the system is a semiconductor. That is why I put Occupation=fixed. I also notice that there is no fractional occupation at any k-point. Then why unoccupied one is lower energy than highest occupied level. Can you help me to rectify the issue here. Attached is out put for the scf calculation from QE.

Regards,

[Bramhachari Khamari]

Dear Sir,
I forgot to attach the scf output from QE in my last post. Here, I attached below,


Regards,

[Daniele Varsano]

Dear Bramhachari,
this means that your system is semimetal, at least according to the DFT functional you are using and you need to consider a smearing in the occupation.

see e.g.:
https://www.mail-archive.com/users@list ... 40212.html

or search in the QE forum as this is often reported.

I also notice that there is no fractional occupation at any k-point.

You cannot have fractional occupation without smearing.


Next please note:

Code: Select all

 kinetic-energy cutoff     =     180.0000  Ry
 charge density cutoff     =    1100.0000  Ry

If you are using norm-conserving PP there is no need to specify charge density cutoff as it is set by default to 4*wf_ecut
see in your output:

Code: Select all

 Message from routine setup:
     no reason to have ecutrho>4*ecutwfc

Best,
Daniele

[Bramhachari Khamari]

Dear Sir,
Thanks a lot for the reply.

Regards,