[Pw_forum] How to calculate magnetization density when SOC is on

Andrea Dal Corso dalcorso at sissa.it
Mon Jan 17 14:42:30 CET 2011


It seems to me that the code is OK. The spin of bands with the same
energy at k and -k is opposite.

If you want only the spin and not the band ordering you can set
no_overlap=.true.
in the bands.x input.

HTH

Andrea 



On Mon, 2011-01-17 at 22:34 +0900, soroush pakseresht wrote:
> Dear Andrea, 
> 
> 
> No, I have no inversion symmetry in my system. So I expect the bands
> at k,up and -k,dn (up,dn:spin) to be the same due to the time reversal
> symmetry.  Let me quote part of my results for spin component 1 (Mx):
> 
> 
> at k
>   0.013  -0.007   0.108  -0.114  -0.108   0.102  -0.047   0.032
>  -0.073   0.094
>    0.250  -0.250  -0.250   0.250  -0.250   0.250   0.243  -0.246
> 0.085  -0.058
>   -0.050  -0.015  -0.058   0.072   0.048  -0.015  -0.075   0.072
> 0.129  -0.110
>   -0.237   0.215  -0.087   0.085  -0.245   0.248  -0.243   0.244
> 0.249  -0.249
> 
> 
> at -k
>   -0.013   0.007   0.114  -0.108  -0.102   0.108  -0.032   0.047
> 0.073  -0.094
>    0.250  -0.250  -0.250   0.250  -0.250   0.250   0.246  -0.243
> 0.058  -0.085
>    0.015   0.050  -0.072   0.058   0.015  -0.048  -0.072   0.075
> 0.110  -0.129
>   -0.215   0.237  -0.085   0.087  -0.248   0.245  -0.244   0.243
> 0.249  -0.249
> 
> 
> I am in particular interested in band #29. As you can see for both k
> and -k they are positive (0.129, 0.110). However, I just found the
> eigen-energies corresponding to band #29 and #30 are just reversed
> between k and -k:
> 
> 
> at k
>  -18.403 -18.403 -18.394 -18.393 -15.427 -15.426 -15.415 -15.415
> -15.401 -15.401
>   -6.240  -6.236  -5.303  -5.292  -3.535  -3.529   1.548   1.611
> 2.796   2.864
>    3.577   3.637   3.964   4.063   4.482   4.540   4.800   4.833
> 7.649   7.776
>    8.465   8.587   9.257   9.264  10.863  10.866  12.514  12.537
>  13.804  13.824
> at -k
>  -18.403 -18.403 -18.393 -18.394 -15.426 -15.427 -15.415 -15.415
> -15.401 -15.401
>   -6.236  -6.240  -5.292  -5.303  -3.529  -3.535   1.611   1.548
> 2.864   2.796
>    3.637   3.577   4.063   3.964   4.540   4.482   4.833   4.800
> 7.776   7.649
>    8.587   8.465   9.264   9.257  10.866  10.863  12.537  12.514
>  13.824  13.804
>  
> As you see, in the former the respective energies of bands #29 and #30
> are  7.649  and  7.776. But for -k, they become 7.776   7.649.
> Apparently the ordering of bands at k and -k are not the same. So I
> guess my mistake was to compare Mx at k with Mx at -k for the same
> band numbers. Do you have any suggestion?
> 
> 
> Thanks,
>  
> 
> 
>  
> On Mon, Jan 17, 2011 at 10:08 PM, Andrea Dal Corso <dalcorso at sissa.it>
> wrote:
>         Please provide more details. Is there inversion symmetry in
>         your system?
>         
>         Andrea
>         
>         
>         On Mon, 2011-01-17 at 21:13 +0900, soroush pakseresht wrote:
>         > Dear Andrea and Gabriele,
>         >
>         >
>         > First, let me thank both of you for your valuable comments
>         and
>         > feedbacks.
>         > As suggested by Andrea, I used bands.x with lsigma=.true. .
>         It seems
>         > to be working. Now I can see the results for Mx, My and Mz
>         components
>         > of magnetization. However, something seems to be wrong. The
>         > magnetization which I get for the same band but at k and -k
>         are
>         > exactly the same. Due to the strong spin-orbit coupling, I
>         am
>         > expecting that the direction of magnetization at k becomes
>         opposite to
>         > that at -k. Is it due to some symmetry constraints, which I
>         might have
>         > not turned off during the SCF and NSCF calculations? Any
>         idea?
>         >
>         >
>         > Looking forward for your response,
>         > Saeed
>         >
>         > On Mon, Jan 17, 2011 at 7:55 PM, Andrea Dal Corso
>         <dalcorso at sissa.it>
>         > wrote:
>         >
>         >         On Mon, 2011-01-17 at 11:48 +0100, Gabriele
>         Sclauzero wrote:
>         >         > Hi Soroush
>         >         >
>         >         > On 01/15/2011 02:50 PM, soroush pakseresht wrote:
>         >         > > Hi everybody,
>         >         > >
>         >         > >
>         >         > > Can anybody tell me how I can calculate the
>         magnetization
>         >         > > M=M(Mx,My,Mz) for a system with spin-orbit
>         coupling.
>         >         >
>         >         > Usually the three components of the total
>         non-collinear
>         >         magnetization
>         >         > are printed at the end of the calculation, e.g.
>         >         >
>         >         >      The total energy is the sum of the following
>         terms:
>         >         >
>         >         >      one-electron contribution =   -90.33299784 Ry
>         >         >      hartree contribution      =    49.63924278 Ry
>         >         >      xc contribution           =   -28.28794844 Ry
>         >         >      ewald contribution        =    -0.25003496 Ry
>         >         >      smearing contrib. (-TS)   =    -0.00000505 Ry
>         >         >
>         >         >      total magnetization       =     0.00     0.00
>         0.11
>         >         Bohr
>         >         > mag/cell
>         >         >      absolute magnetization    =     0.11 Bohr
>         mag/cell
>         >         >
>         >         >
>         >         >
>         >         > > I'd like to know the total value of Mx, My, Mz
>         for each
>         >         band at a
>         >         > > given kpoint.
>         >         > >
>         >         >
>         >
>         >
>         >         You can use the program bands.x with the flag
>         lsigma=.true. .
>         >         Please check INPUT_BANDS for more info.
>         >
>         >         HTH
>         >
>         >         Andrea
>         >
>         >
>         >
>         >         > What you're asking for is somehow the
>         magnetization carried
>         >         by the
>         >         > individual KS eigenstates. I think that for this
>         you'll have
>         >         to modify
>         >         > the PWscf or PostProc code. For instance you could
>         start
>         >         from pp.x and
>         >         > modify the option plot_num=7 by performing the
>         integral of
>         >         the
>         >         > resulting spin-density over the whole unit cell
>         (should be
>         >         the sum
>         >         > over the FFT components, perhaps multiplied by
>         some
>         >         normalization
>         >         > factor).
>         >         > Maybe it's easier to implement this directly in
>         PWscf. At
>         >         the moment
>         >         > the code first computes the total spin-density
>         (which is the
>         >         sum over
>         >         > bands and k-points of the individual KS densities)
>         in
>         >         sum_band.f90
>         >         > (search for 'domag') and then the magnetization by
>         >         integrating the
>         >         > components of the spin-density (see
>         compute_magnetization in
>         >         > PW/electrons.f90). You can modify sum_band.f90 by
>         adding a
>         >         call to a
>         >         > function analogous to compute_magnetization but
>         for each
>         >         single KS
>         >         > state.
>         >         >
>         >         >
>         >         > HTH
>         >         >
>         >         > GS
>         >         >
>         >         > >
>         >         > >
>         >         > > Thanks
>         >         > >
>         >         > > --
>         >         > > Soroush Pakseresht
>         >         > > Advanced Science Institute (RIKEN)
>         >         > > Saitama-Japan
>         >         > >
>         >         > >
>         >         > >
>         >         > > _______________________________________________
>         >         > > Pw_forum mailing list
>         >         > > Pw_forum at pwscf.org
>         >         > >
>         http://www.democritos.it/mailman/listinfo/pw_forum
>         >         > >
>         >         >
>         >         >
>         >         > --
>         >         >
>         >         > Gabriele Sclauzero, EPFL SB ITP CSEA
>         >         > PH H2 462, Station 3, CH-1015 Lausanne
>         >         > _______________________________________________
>         >         > Pw_forum mailing list
>         >         > Pw_forum at pwscf.org
>         >         > http://www.democritos.it/mailman/listinfo/pw_forum
>         >         --
>         >
>         >         Andrea Dal Corso                    Tel.
>         0039-040-3787428
>         >         SISSA, Via Bonomea 265              Fax.
>         0039-040-3787249
>         >         I-34136 Trieste (Italy)             e-mail:
>         dalcorso at sissa.it
>         >
>         >
>         >
>         >         _______________________________________________
>         >         Pw_forum mailing list
>         >         Pw_forum at pwscf.org
>         >         http://www.democritos.it/mailman/listinfo/pw_forum
>         >
>         >
>         >
>         >
>         > --
>         > Soroush Pakseresht
>         > Advanced Science Institute (RIKEN)
>         > Saitama-Japan
>         >
>         >
>         > _______________________________________________
>         > Pw_forum mailing list
>         > Pw_forum at pwscf.org
>         > http://www.democritos.it/mailman/listinfo/pw_forum
>         --
>         Andrea Dal Corso                    Tel. 0039-040-3787428
>         SISSA, Via Bonomea 265              Fax. 0039-040-3787249
>         I-34136 Trieste (Italy)             e-mail: dalcorso at sissa.it
>         
>         
>         _______________________________________________
>         Pw_forum mailing list
>         Pw_forum at pwscf.org
>         http://www.democritos.it/mailman/listinfo/pw_forum
>         
> 
> 
> 
> -- 
> Soroush Pakseresht
> Advanced Science Institute (RIKEN)
> Saitama-Japan
> 
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
-- 
Andrea Dal Corso                    Tel. 0039-040-3787428
SISSA, Via Bonomea 265              Fax. 0039-040-3787249
I-34136 Trieste (Italy)             e-mail: dalcorso at sissa.it





More information about the users mailing list