Metallic system

[lamia]

Dear developers,

I am trying to use Yambo in order to improve the DFT's band gap of my system.
I have this problem that Yambo sees my system as Metallic, while the band gap is 0.4 eV.

[WARNING]Metallic system

How could I get band gap correction when Yambo sees my system as metallic?!

[Daniele Varsano]

Dear Lamia,

the system is seen as metallic by yambo or it is already metallic in the QE ground state calculation?
If it is not the case, can you post your QE output and r_setup?

Best,
Daniele

[lamia]

Dear Daniele,

Thank you for your reply.
The attached is the output and r_setup files.

Best,

Lamia

[Daniele Varsano]

Dear Lamia,

some suggestions:
if the system is a semiconductor, try to run QE using the occupation fixed (no smearing).
Next, yambo cannot perform a GW calculation not because of the metallicity but because you have a non-uniform grid:

Code: Select all

[WARNING][RL indx] 2 equivalent points in the rlu grid found

 [RL indx] X grid is not uniform. Gamma point only.
 [WR./SAVE//ndb.kindx]---------------------------------------
  Fragmentation                    :no
  Polarization last K   :  90
  QP states             :   1   90
  X grid is uniform     :no
  BS scattering         :no
  COLL scattering       :no

In order to overcame this problem first have a look to the QE messages:

Code: Select all

Message from routine setup:
     no reason to have ecutrho>4*ecutwfc
     Message from routine setup:
     using ibrav=0 with symmetry is DISCOURAGED, use correct ibrav instead

Try to use the correct ibrav.

If this not solve the problem in the yambo setup, you can add this flag in the yambo setup input:

Code: Select all

NoDiagSC

Best,

Daniele

[lamia]

Dear Daniele,

The system that I'm working on is a semiconductor. It doesn't work even with the suggestions. Any other solutions, please ?




Best,

Lamia

[Daniele Varsano]

Dear Lamia,

I think you should solve first the problem of the non-uniform k point grid.
This should be achieved by defining an appropriate ibrav and an automatic k point sampling (possibly centred at gamma, which is not now the case).

Best,
Daniele