Menu

Concerning spin-spiral calculations

Elk Users
lwollmann
2015-01-19
2015-01-20
  • lwollmann

    lwollmann - 2015-01-19

    Dear developers and users,

    I'm writting to you concerning some understanding issues on my side while conducting spin-spiral calculations. Especially the spin-spiral groundstate of gamma-Iron, given as an example lead to some confusion

    The given elk.in is clear except the applied field through out the unit cell "bfieldc", as the direction under investigation is supposed to be [qq0] with a fixed q of [0.1,0.1,0] for the given example.

    ! spin-spiral q-vector in lattice coordinates
    vqlss
    0.1 0.1 0.0
    
    ! spiral magnetic field along q with perpendicular component only
    bfieldc
    0.05 0.00 0.0
    

    Condensing this to one question as: "Why isnt the field applied along q with two finite components (b_x,b_y,0) ?"

    I'd be happy to be enlightened by an expert and thanks in advance.
    Lukas

     

    Last edit: lwollmann 2015-01-19
  • Lars Nordström

    Lars Nordström - 2015-01-20

    Dear Lukas,

    Since spin-orbit coupling cannot be included in a spin spiral calculation,
    the real/reciprocal space and spin space are not coupled, i.e. they are independent of each other.

    A spin spiral is then taken to always rotate around the z-axis in spin space in ELK. For instance a planar spin spiral is in the xy-plane in spin space, as specified by bfieldc in your case. The exact direction within this plane is irrelevant since it is only specifying the direction of one atomic site within an infinite spin spiral ...

    However, the wave-vector can point in any direction in reciprocal space - as specified by vqlss.

    In short the answer is that the two vectors are completely independent.

    Good luck,
    Lars

     

Log in to post a comment.