Problem with yambo in considering non-metallic sysytems

[rvatanme]

Dear all,

I'm trying to apply GW approximation on DFT calculations obtained from quantum espresso (QE). The system that I'm working on is an insulator, but Yambo is considering it as a metallic system even with "occupations = fixed" option in QE input. Could any one please help me to fix this problem.

The attached is my input and r_setup files.

Thanks in advances,
Reza

[Daniele Varsano]

Dear Reza,
most probably it happens because the energy gap as calculated by QE is very small (check it). Yambo recalculate occupations and see the system as metallic.
If this is the case you can generate the setup input with high verbosity:

Code: Select all

yambo -i -V all

and reduce the default value for the occupation threshold:

Code: Select all

OccTresh=0.1000E-4           # Occupation treshold (metallic bands)

or just insert this variable by hand in your input file.

Best,
Daniele

[rvatanme]

Dear Daniele,

It didn't work. Thanks anyway for replying.

Best,
Reza

[Daniele Varsano]

Dear Reza,
I will try to reproduce your problem.
Best,
Daniele

[rvatanme]

Dear Daniele,

Thank you so much for your kind.

Best,
Reza

[Daniele Varsano]

Dear Reza,
I've run your qe input, everything converge smoothly, but when running the "nscf" calculation at the end QE says:

Code: Select all

highest occupied, lowest unoccupied level (ev):     6.2657    6.2276

As you can see, the lowest unoccupied state is lower in energy than the highest occupied, indication that the system is "metallic", and yambo indeed see it as metallic.
You can have a look to this thread in QE forum.

If you are sure, for some reason (literature etc.) that your system should be semiconductor also at DFT /BLYP level check your input file (structure and other parameters etc.). Consider you are using a very low energy cutoff, where hgh pseudopotentials are quite hard and do require a large energy cutoff (even if this will not solve your problem). Consider also to update your yambo version, you are using a quite old one, and many bugs have been fixed in the new release.


Best,
Daniele