Nonzero temperature from QE database

[Weiqing Zhou]

Hi all,

After initialization (yambo), I monitor the report file r_setup and find that the temperature from QE output sometimes is nonzero, always 0.02585 eV (300K)

[RD./SAVE//ns.db1]------------------------------------------
Bands : 150
K-points : 1214
G-vectors [RL space]: 67815
Components [wavefunctions]: 8548
Symmetries [spatial]: 4
Spinor components : 1
Spin polarizations : 1
Temperature [ev]: 0.02585
Electrons : 28.00000
WF G-vectors : 10323
Max atoms/species : 2
No. of atom species : 3

-----------------------------------------------------------------------
And in this case, Default of ElecTemp in all following yambo input will be 0.025 eV, too. I think this change maybe make some difference in my calculation
compared to the 0K one.
I also find nonzero temperature only occur when I set ' smearing = 'mp' ' in PWscf.
Why the temperature of the DFT ground state is not zero ? Does it matter ?

Weiqing

[Daniele Varsano]

Dear Weiqing,
if you use a smearing then you will have a finite electronic temperature (it affects the occupation numbers). In presence of a gap this should not affect your results. In any case, if your system is not metallic, you can set 0K by adding in your input files (also for the setup) the variable:

Code: Select all

ElecTemp= 0.0 eV

You can also set an arbitrary temperature in case of a metallic system but 0K it is not allowed.

Best,
Daniele