Page 1 of 1

ISMEAR results in different spin magnetic moments

Posted: Fri Feb 13, 2015 2:26 am
by leegeunsik
Hi all,

I want to ask about a strange result that I am experiencing.
The system is P-doped MoS2 3x3 monolayer.
I expect non-zero magnetic moment.

But with following tags in INCAR and sufficient k-points, I got "mag=0.0".

ISPIN = 2
ISMEAR = 0
SIGMA = 0.1

1 F= -.19469684E+03 E0= -.19466101E+03 d E =-.716636E-01 mag= 0.0015

It is also zero when I use ISMEAR=-1.

However when I use ISMEAR=1, I get following nonzero value.
1 F= -.19469530E+03 E0= -.19469295E+03 d E =-.704062E-02 mag= 0.7200
Also when ISMEAR=-5, it is non-zero as following.
1 F= -.19467525E+03 E0= -.19467525E+03 d E =0.000000E+00 mag= 0.9773

As I checked SIGMA and k-grid, it does not affect the result above.

Would you please explain why the result depends qualitatively on the smearing method?

With best regards,
Geunsik Lee

Re: ISMEAR results in different spin magnetic moments

Posted: Mon Feb 16, 2015 9:05 am
by admin
Smearing is a numerical tool for improving the convergency of the calculation.
At increased smearing electrons are allowed to go into bands which should stay empty.
This way smearing influences the final magnetic state.
Check occupancies of bands of both, spin-up and spin-down.
In pure magnetic state only integer occupancies are allowed.