wrong site resolved magnetic moments

Elk Users
semran
2012-12-17
2013-06-11
  • semran
    semran
    2012-12-17

    Hi all,

    I have trying to produce the site resolved spin magnetic moments of Mn2CoAl. Space group of the structure is 216 and it is an inverted Heusler. The site resolved spin magnetic moments for Mn and Co are as follows:

    Mn(4a):-2.6mb
    Mn(4c):1.6 mb
    Co=-1 mb

    Unfortunately, these value do not agree with the literature. They should be:
    Mn (4a)=-2 mb
    Mn (4c)=3 mb
    Co=1 mb

    Here is my input file:
    *********************************************************************************************************************
    Heuslar Mn2CoAl with conventional unitcell taken from spacegroup utility. converged parameters have been used and vkloff for fcc structure has been used.
    ! Note the large global magnetic field that is
    ! halved at every iteration. This procedure speeds up the convergence of
    ! magnetic calculations. Compare the number of iterations with the Fe example.
    !tasks 0 for single point calculations

    tasks
      0

    spinorb
    .true.

    spinpol
    .true.

    ! large magnetic field in the z-direction
    bfieldc
      0.0  0.0  4.0
    ! the magnetic field is halved at every iteration
    reducebf
      0.5

    xctype
    20

    swidth
    0.001

    primcell
    .true.

    trimvg
    .true.

    tforce
    .true.

    maxscl
    400

    beta0
    0.01

    rgkmax
    10.0

    gmaxvr
    15.0

    ! Pulay mixing with spin-polarized calculation
    ! is effective only if a large field is provided,
    ! such that the system does not get stuck in a non-magnetic solution
    mixtype
      1

    ! fairly large number of empty states required for magnetic cases
    nempty
      100

    avec
       10.95000000       0.000000000       0.000000000
       0.000000000       10.95000000       0.000000000
       0.000000000       0.000000000       10.95000000

    atoms
       3                                    : nspecies
    'Al.in'                                 : spfname
       4                                    : natoms; atposl, bfcmt below
        0.75000000    0.75000000    0.75000000    0.00000000  0.00000000  0.00000000
        0.75000000    0.25000000    0.25000000    0.00000000  0.00000000  0.00000000
        0.25000000    0.75000000    0.25000000    0.00000000  0.00000000  0.00000000
        0.25000000    0.25000000    0.75000000    0.00000000  0.00000000  0.00000000
    'Co.in'                                 : spfname
       4                                    : natoms; atposl, bfcmt below
        0.50000000    0.50000000    0.50000000    0.00000000  0.00000000  0.00000000
        0.50000000    0.00000000    0.00000000    0.00000000  0.00000000  0.00000000
        0.00000000    0.50000000    0.00000000    0.00000000  0.00000000  0.00000000
        0.00000000    0.00000000    0.50000000    0.00000000  0.00000000  0.00000000
    'Mn.in'                                 : spfname
       8                                    : natoms; atposl, bfcmt below
        0.25000000    0.25000000    0.25000000    0.00000000  0.00000000  0.00000000
        0.25000000    0.75000000    0.75000000    0.00000000  0.00000000  0.00000000
        0.75000000    0.25000000    0.75000000    0.00000000  0.00000000  0.00000000
        0.75000000    0.75000000    0.25000000    0.00000000  0.00000000  0.00000000
        0.00000000    0.00000000    0.00000000    0.00000000  0.00000000  0.00000000
        0.00000000    0.50000000    0.50000000    0.00000000  0.00000000  0.00000000
        0.50000000    0.00000000    0.50000000    0.00000000  0.00000000  0.00000000
        0.50000000    0.50000000    0.00000000    0.00000000  0.00000000  0.00000000

    ngridk
    19  19  19

    **********************************************************************************************************************

    Thanks all.
    best regards,

    ipek

     
  • Markus
    Markus
    2013-01-03

    Hi ipek,

    your results are fine. Eventually your lattice constant is a bit small. Your nempty and rgkmax are overkill. nempty = 40 is more than sufficient, rgkmax = 8.0 is certainly enough.

    Which reference do you compare these values to?

    Best regards,
    Markus

     
  • semran
    semran
    2013-01-04

    Hi Markus,

    I really appreciate for your help. The paper which I am trying to produce the results is:"Realization of spin gapless semiconductors:the heusler compound Mn2CoAl"-Claudia Felser et al.

    the paper says that magnetic moments of Mn(4a)=-2 mb, Co(4b)=+1mb, Mn(4c)=+3mb.

    But I have found that Mn(4a):-2.6mb Mn(4c):1.6 mb Co=-1 mb . As you see Co and Mn(4c) dont seem to have right moments. However I can confirm the antiferromagnetic coupling between neighboring Mn sites, Co atom does not have the right  moment. I found it -1mb but it should be +1 mb.
    I will make a large unit cell. And share the results with you. If you have any other ideas, please could you guide me?

    Thanks again.

    BTW, Happy new year all Elk users and developers.

    regards,

    semran

     
  • Markus
    Markus
    2013-01-04

    Hi Semran,

    the paper says "approximately". Have a look at my article J. Phys. Condens. Matter 23 (2011) 116005. I've calculated some properties of these compounds and found pretty much the same values with SPRKKR as you did. I've also checked the numbers with Elk and they almost agree (Elk's values are a bit higher). In Ouardi's article they use ASW, which may give different results.

    By the way, remember that the muffin-tin moments depend on the radii you choose. Try to increase or decrease the radii in the species files to see what I mean.

    Considering the signs: You have a positive bfieldc, so your moments are negative (see the total moment!). Reverse the sign of bfieldc to get a positive total moment. That's more intuitive.

    Here's a simplified elk.in for Mn2CoAl using a different convention for the basis vectors to prevent Elk from changing the origin: Co (4a), Mn (4b), Mn (4c), Al (4d). The lattice constant is from Liu et al., Phys Rev. B 77, 104424 (2008).

    tasks
     0
    spinpol
     .true.
    xctype
      20
    bfieldc
      0.0 0.0 -2.0
    reducebf
     0.85
    sppath
      '/home/meinert/elk/elk-1.4.22/species/'
    ngridk
     15 15 15
    primcell
    .true.
    nempty
     40
    swidth
     0.001
    scale
      11.04
    avec
       0.5  0.5  0.0
       0.0  0.5  0.5
       0.5  0.0  0.5
    atoms
       3
    'Mn.in'
       2
        0.25  0.25  0.25
        0.50  0.50  0.50
    'Co.in'
       1
        0.00  0.00  0.00
    'Al.in'
       1
        0.75  0.75  0.75
    

    With this elk.in and the default species files I find:
    Co 1.05 m_B, Mn 2.67 m_B, Mn -1.7 Mn. Co and the nearest-neighbor Mn have parallel moments.

    Regards,
    Markus

     
  • semran
    semran
    2013-01-04

    Hi Markus,

    Actually, I have already read your paper entitled "Itınerant and localized Mn moments in ferrimgnetic Mn2CoGa thin film.." Your paper is also one of my guidance during calculations. I will check the muffin-tin radius again and reverse the magnetic field on the system. I will share the results with you. I really apprecate for your time and kindly helps.

    best regards,

    S.ipek