[Wien] Fermi Level

Stefaan Cottenier Stefaan.Cottenier at UGent.be
Tue Dec 2 11:52:32 CET 2014


> I did a scf calculation for 1000 k-points and I got a fermi level. (for
> example, Ef=0.41000)
> Then I run x kgen for 150000 k-points and I got a new fermi level. (for
> example, Ef=0.48000)
> Which fermi level should I use in the case.intrans file for transport
> properties?

Neither of both (yet). You must test how your properties of interest 
(i.c. the fermi level) behave as a function of the number of k-points. 
Only then you can decide how many k-points you need in order to have the 
precision you want. Perhaps with 300000 k-points the fermi level will be 
0.52 (unlikely, yet just as an example), or perhaps with only 20000 it 
will have settled at 0.48 and stay there whichever many k-points you add.

Stefaan



More information about the Wien mailing list