Why band structure & fermi energy is changing with fermi broadening temperature?

[rounoknaimniloy]

I was calculating band structure of nanoribbon a material. I have learned from the forum that 'Fermi Broadening temperature ' has no physical significant. I calculated band structure for different Broadening temperature.  I observed that fermi energy is shifting downwards. And there is significant changes in the band structure also. At 1800K I got dirac cone at gamma point but occupied states above the fermi energy. But at 4200k I got fermi level right with10meV gap which is not what I want. Can anyone please explain this to me?

[Tue Gunst]

Hi,
The Fermi-broadening does have a significant impact on DFT results.
Specifically, it impacts how fast scf converges and the final state (if the broadening is too large).
For a semiconductor I would reduce the smearing to 100-300 K. This means that more k-points are needed.
For a metal a larger broadening can be used and lower k-points sampling.
In any case, the broadening should be reduced and k-point sampling increased until simulations are converged.
In general the defaults are good but you seem to use a much larger broadening.