[Pw_forum] big difference between "total magnetization" and "absolute magnetization" for 4f elements

[琨陶]

Thank you for your suggestions. When I set two atoms in the unit cell
to be antiferromagnetic, the "total magnetization" is nearly zero, and
the "absolute magnetization" is about 14.67. And the magnetic moment
for two atoms are about 7.16 and -7.18, which is more or less
acceptable.

Now, the question is that why "total magnetization" and "absolute
magnetization" differs so much if I want to force two atoms into a
ferromagnetic configuration.

Any suggestions? Thanks!

With regards,
Kun Tao


On 13 March 2013 11:14, Andrei Malashevich <andrei.malashevich at yale.edu> wrote:
> If you want to model antiferromagnetic structure, you should define two
> types of atoms, Gd1 and Gd2.
> Change ntyp to 2. Use, e.g.,
>  starting_magnetization(1)=1.0
> starting_magnetization(2)=-1.0
>
> ATOMIC_SPECIES
>  Gd1  1.0  Gd.GGA-PBE-paw.UPF
>  Gd2  1.0  Gd.GGA-PBE-paw.UPF
> ATOMIC_POSITIONS {crystal}
> Gd1  0.33333333  0.66666667  0.25
> Gd2 0.66666667  0.33333333  0.75
>
>
> On Wed, Mar 13, 2013 at 11:05 AM, 琨陶 <taokun76 at gmail.com> wrote:
>>
>> Hi,
>>
>> Yes, the bulk Gd without Hubbard U is antiferromagnetic. If I set the
>> two atoms in the unit cell to be antiferromagnetic, the "total
>> magnetization" should be equal to zero, and the "absolute
>> magnetization" is what I have gotten. Now, however, the "total
>> magnetization" is not zero. Could you give me some suggestions?
>> Thanks!
>>
>> With regards,
>> Kun Tao
>>
>> On 13 March 2013 10:48, Stefano de Gironcoli <degironc at sissa.it> wrote:
>> > are you sure you are defining the correct magnetic structure ?
>> > is Gd maybe antiferromagnetic ?
>> > stefano
>> >
>> > On 03/13/2013 03:19 PM, 琨陶 wrote:
>> >> Dear everyone,
>> >>
>> >> I want to calculate the magnetic properties of bulk Gd whose magnetic
>> >> moment is about 8 Bohr magneto/atom. Usually, in the output file
>> >> values of "total magnetization" and "absolute magnetization" should be
>> >> nearly the same. However, in my output file, I found a big difference
>> >> between them as following:
>> >>
>> >>       total magnetization       =     8.00 Bohr mag/cell
>> >>       absolute magnetization    =    15.72 Bohr mag/cell
>> >>
>> >> The good news is that the "absolute magnetization" is close to the
>> >> experimental results (7.86 Bohr mag/atom), the bad news is that the
>> >> "total magnetization" is wrong. Moreover, when I try to calculate the
>> >> magnetic moment on each atom with projwfc.x program, I got about 4.1
>> >> Bohr mag/atom.
>> >>
>> >> Could you give me some suggestions? Any advice are appreciated, thanks!
>> >>
>> >> Regards,
>> >> Kun Tao
>> >>
>> >>
>> >> Attached below is the input file:
>> >>
>> >>   &control
>> >>      calculation='scf'
>> >>      restart_mode='from_scratch',
>> >>      prefix='Gd',
>> >>      pseudo_dir = './'
>> >>      outdir='./',
>> >>   /
>> >>   &system
>> >>      ibrav= 4, celldm(1)= 6.96, celldm(3)= 1.59,
>> >>      nat= 2, ntyp=1,
>> >>      nspin = 2,
>> >>      starting_magnetization(1)=1.0,
>> >>      ecutwfc=55.0,
>> >>      ecutrho=220.0,
>> >>      occupations ='smearing',
>> >>      smearing ='gauss',
>> >>      degauss = 0.01,
>> >>   /
>> >>   &electrons
>> >>      diagonalization='david',
>> >>      mixing_beta = 0.3,
>> >>      conv_thr = 1.0d-8,
>> >>   /
>> >> ATOMIC_SPECIES
>> >>   Gd  1.0  Gd.GGA-PBE-paw.UPF
>> >> ATOMIC_POSITIONS {crystal}
>> >> Gd  0.33333333  0.66666667  0.25
>> >> Gd  0.66666667  0.33333333  0.75
>> >> K_POINTS {automatic}
>> >> 4 4 4 0 0 0
>> >> _______________________________________________
>> >> Pw_forum mailing list
>> >> Pw_forum at pwscf.org
>> >> http://pwscf.org/mailman/listinfo/pw_forum
>> >
>> > _______________________________________________
>> > Pw_forum mailing list
>> > Pw_forum at pwscf.org
>> > http://pwscf.org/mailman/listinfo/pw_forum
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://pwscf.org/mailman/listinfo/pw_forum
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum