ABINIT Discussion Forums

Hello everyone:

In order to calculate l-resolved local DoS, we should set up prtdos=3

and specify

natsph, iatsph and ratsph.

I'm sorry that I still cant understand the meaning after looking at the input-variable-explanation, tutorial as well as trying...

For instance, if there are 3 atoms in one unit cell, and now the local DoS of each of the three atoms are needed,

Then shall we setup as
ndtset 3
natsph 1
iatsph1 1
iatsph2 2
iatsph3 3
ratsph1 r1
ratsph2 r2
ratsph3 r3

Or natsph 1 iatsph 1 2 3 ratsph r1 r2 r3?
It seems none works...

What's worse, I don't quite follow why iatsph <=natsph, since the tutorial example of FCC Fe in spin session, it does have iatsph>natsphe...

Any hint of this? Many thanks.

Hi,

natsph and iatsph work just like ndtset and jdtset.

natsph = the number of atom you want your decomposition in (default = natom)
iatsph = the index of these atom (must be smaller or equal to natom; you have to specify natsph of these) (default= 1 to natom)

In your case, you could do it in 1 dataset with natsph =3 and iatsph = 1 2 3 (or not specify anything, it's the same by default).

Simon

Thank you very much, Simon! Your answer exactly sweeps away my confusion and helps a lot!