[X] Metallic system

[Laura Caputo]

Hello,

I have calculated the WFK.nc file via Abinit and then converted it to Yambo with a2y. When I initialize the job, however, I obtained the '[WARNING][X] Metallic system' message.

Reading previous posts, I have modified the input file adding:
setup # [R INI] Initialization
Nelectro= 120.00000 # Electrons number
ElecTemp= 0.000000 eV # Electronic Temperature
BoseTemp=-1.000000 eV # Bosonic Temperature
OccTresh=0 # Occupation treshold (metallic bands)

What can I do to get rid of this message? Is this smearing related?
In attachment the report file (I tried to upload also the output file for abinit but it's too large).

I add also that the gap at DFT level is quite low (0.047eV).

Sincerely,
Laura Caputo

[Daniele Varsano]

Dear Laura,
Yambo calculates the Fermi level and it sees the system as metallic, even if you set ElecTemp=0
You can compare the occupation numbers shown in the report with the ones of abinit and see what is going on.
Also, yes are you able to converge the ground state without using a smearing?

Best,
Daniele

[Laura Caputo]

Dear Daniele,

Looking at Abinit output, the Fermi level calculated by it
Fermi (or HOMO) energy (hartree) = -0.03152 Average Vxc (hartree)= -0.15716
is different from Yambo. Is this normal?

I have already calculated it without smearing actually, I checked it.
I have attached the output compressed. From that, I can see that 60 bands are occupied (as expected) while 61 are empty. On the contrary in the report the filled bands are 58, from 59 to 61 are seen as metallic and then are empty.

[Daniele Varsano]

Dear Laura,

Fermi (or HOMO) energy (hartree) = -0.03152 is different from Yambo. Is this normal?

Yes, as Yambo recalculate the Fermi energy. In case of a semiconductor the HOMO (or VBM) is set to zero.

I'm not practical with the abinit output, it is more than 10 years I do not make a calculation using abinit so maybe I'm wrong in reading your output.

Anyway, it seems to me it is a relax calculation and the Fermi energy of the last iteration is -0.03071 Ha.
The eigenvalues (and occupations) are given only for 1 kpt: kpt# 1, nband= 64, wtk= 0.01852, kpt= 0.0000 0.0000 0.0000 so it is not easy to compare the electronic structures of abinit and what is read from yambo.
I do not know how, but it should be a way to look at the eigenvalues for each one of the kpoints and see how the ground state looks like.

Best,
Daniele

[Laura Caputo]

Dear Daniele,

I have only another question. I have seen that on Quantum Espresso one could 'reinforce' the ground state calculation for non metallic systems with occupations="fixed".
Could it be useful to try this if the problem persists? Since I understand that the yambo results should be independent of the code used for the ground state calculation.

[Daniele Varsano]

Dear Laura,
yes you can try to use QE.
occupations="fixed" will force the system to be semiconductor and it will not converge in case it is not.

Best,
Daniele