[QE-users] [QE-users ]Opposite HOMO-LUMO energies but no "System is metallic, smearing is needed"

[Konrad Gruszka]

Dear community

Recently I've made some calculations of Heusler compound using 
occupations='fixed'.

Until now, I was quite sure, that if I incorrectly choose occupations to 
'fixed', QE will show me known "System is metallic, smearing is needed" 
error and then interrupt my calculation. Now, after I made my 
calculation with occupations='fixed', I got:

"highest occupied, lowest unoccupied level (ev):     4.8994 3.9581"

which I interpret as metal, yet no "smearing needed" error occur.

In addition, in the XML files made after SCF/NSCF calculations show 
totally different FERMI/HOMO/LUMO close to about 0.177eV/0.18/0.145.

The HOMO/LUMO from SCF is  therefore not consistent with DOS plot (which 
suggests HOMO - LUMO gap of about 5 eV)  and BAND structure (which I 
think take FERMI from XML file, and shows that system is metallic).

This behavior I see for 6.7, 6.6 and 6.2.1 versions of QE (in terms of 
SCF HOMO/LUMO levels and no error about metallic state).

What is the reason for such behavior? Which values can I trust in the 
newest 6.7 version? Is there no more "system is metallic" message? Or is 
my interpretation of reverse HOMO-LUMO not correct?


Konrad Gruszka