[Pw_forum] electron-phonon and nspin=2

[Andrea Floris]

Dear users,
I have a question about el-ph calculation in magnetic systems (nspin). In 
PH/phq_readin.f90, there is
the line

IF (elph.AND.lsda) CALL errore ('phq_readin', 'El-ph and spin not &
implemented', 1)

that stops a el-ph calculation when nspin=2.. But having a look at the 
file PH/elphon.f90 (containing the routines elphon, elphel and elphsum)
I think that the current implementation allows for such a calculation 
(unless really minor modifications are needed).
In fact in the wavefunctions  the spin is attached to the k point, and the 
scf perturbed potential dvscfin (i.e. dV_Hxc(is)) is always
spin resolved. So the elph matrix-elements will be spin dependent through 
the k point, coupling always k and (k+q) states with the same spin.
Of course one can always comment out the line, but my  question is: Is 
there a particular reason for this call in  PH/phq_readin.f90?
Or is it just because k-summed quantities like total \lambda or the 
linewiths will not be printed in a spin resolved way?

Thank for your answer,
Andrea


---------------
  Dr. Andrea Floris
  Institut fuer Theoretische Physik
  Freie Universitaet Berlin
  Arnimallee 14
  D-14195 Berlin
  Germany
  e-mail: andrea.floris at fu-berlin.de