Spin-polarized USPEX calculations using VASP

[Sylvian Cadars]

Hi,

I very recently started to use USPEX with VASP and I am wondering whether it was possible to set initial magnetic moments in USPEX calculations using variable compositions. I read somewhere in the manual that is was possible to work with multiples of a certain composition, for example in SiO2, to work with compositions x*SiO2 with x ranging for example from 1 to 4. 

Is it possible in a case like this to use spin-polarized calculations and to set an initial magnetic moment to some of the atoms My understanding is that in many systems VASP will not find the correct magnetic phase unless correct initial magnetic moments are provided with the MAGMOM option. The problem is that the magnetic moment must be provided for all atoms in the cell in the initial INCAR file (Specific/INCAR_1). Is there a way to deal with this in USPEX?

This question is in fact very much related to another, which was posted in 2014 with subject "Magnetic Order" but has apparently not been answered... I attach this question below for convenience:

Thanks for your help.
Best regards.


************************
Hi All,

I plan to use VASP along with USPEX and I was wondering if USPEX can
predict the correct magnetic order of ground state structures?

I think this might be more something to do from the VASP side rather
than from the USPEX side. I would like to know how USPEX associates
each chemical specie of the INPUT.txt file with a magnetic moment, if
in my INCAR file I set the tag MAGMOM (+/- 1) for each atom?

As academic example, how do I generate a Ni anti-ferromagnetic and Ni
ferromagnetic phase with USPEX? If I want a Ni anti-ferromagnetic
phase, how in the INPUT.txt file I can make the difference between the
same chemical specie (here Ni) that has a different spin orientation?
Would it be like dealing with two different chemical species? It would
be really great if you could give me a practical example of input
files for USPEX and VASP?

Many Thanks,
Michel.
****************************

[Denise Adorno]

 Dear Sylvian Cadars,


I faced the same problem and I solved by including these following lines in my submitJob_local.m file:

%Insert the MAGMOM tag in INCAR for every structure
[c,d]=unix(['sed -n 6p POSCAR']);
f=str2num(d);
Nb=f(1);
X=num2str(Nb);

fileID=fopen('INCAR', 'a');
fprintf(fileID, 'MAGMOM= %s*2 100*0\n', X);
fclose(fileID);


form my case I set the MAGMOM valuer as 2 for my first atom.

I hope this help you with your case.

Cheers

Denise Adorno
KTH - Reactor Physics