[Sax-project-users] fine doing

[marsamos at democritos.it]

Dear Ali, GW is now a field where there still does not exist enough 
benchmark calculation to a-priori know the convergence parameters. So 
yaou have to perform convergence for k points, cutoff polarizability, 
number of bands.

Cutoff fock is enough to put the same cutoff as for wave functions
for cutoff wave functions and k points first converge DFT band 
structure, and use this parameters for GW.
for the number of bands in a bulk for having gap converges within 0.1 
eV is enough to use 4*times the number of valence states. For a 
molecule, you need 20*times (in the worst case, which is a molecule 
with an electron-affinity of the order of eV).

for the cutoff polarizability usually 6 Ry is enough, in some systems 
you need more (never up to 12Ry), in some systems less. First start 
with 2Ry, then test 4 Ry, 6 Ry amd 8 Ry if the variations are within 
0.1 eV is OK.

bests

Layla

Quoting kazempoor at ph.iut.ac.ir:

>
> Hi ALL
> I am using sax to correct the band gap of some semiconductors. But I 
> don't know how can I do fine calculation w.r.t to optimized 
> parameters.
> for example How can I choose the kpoints mesh? shall I converge 
> them(sure !)? what about cutoff polarizability and cutoff _fock?
>
> In general how can I find that I have done best calculation for GW?
> thanks a lot
> Ali Kazempour
> Physics Department, Isfahn University of Technology
> _______________________________________________
> Sax-project-users mailing list
> Sax-project-users at qe-forge.org
> http://qe-forge.org/mailman/listinfo/sax-project-users
>



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.