[Wien] Question on spin scaling in LDA

[tran at theochem.tuwien.ac.at]

Hi,

It seems that there is a mistake for scaling with LDA.
In vxclm2.f (in directory SRC_lapw0) there are these
three lines:
           vxcu1=vxu
           vxu=(vxcu1*(1.d0+XC1)+vxd*(1.d0-XC1))/2.d0
           vxc=(vxcu1*(1.d0-XC1)+vxd*(1.d0+XC1))/2.d0
in the 3rd line it should be "vxd" instead of "vxc":
           vxd=(vxcu1*(1.d0-XC1)+vxd*(1.d0+XC1))/2.d0

Probably there is the same error in the WIEN2k that you
are using. So, you need to replace vxc by vxd and recompile lapw0.

F. Tran


On Wed, 30 Mar 2016, Wen Fong Goh wrote:

> Hi,
> 
> I am using the spin scaling option to reduce the spin polarization in an antiferromagnetic calculation. I tried xc1 = 0 in both LDA and GGA, but only GGA gives
> zero magnetic moment. We know that with xc1 = 0, the magnetic part of the exchange correlation functional vanishes (equation 11 in Ortenzi et al. 2012), so the
> system should become non-magnetic. Do you know what is going on with the spin scaling in LDA?
> 
>