[Wien] Magnetic dipole term and Orbital polarization

[gimbert at cemes.fr]

Dear Pavel Novak, dear Peter Blaha,

I thank you for your replies to my message.

Concerning the difference between spin moments in scf and scfdmup files:
indeed, the difference is due to the contribution of orbitals s and p, thank
you for this information.

Concerning the magnetic dipole term Tz: I have always a problem with that. For
the radial part, I used RINDEX = 1 (which corresponds to a constant radial
part) and LSINDEX = 8 as you have told me, but the result which I obtained
for bulk hcp Co is high, while it should be nearly zero: I obtained 3.3
instead of 0.014 (Wu, Freeman, J. Magn. M. Mat. 200, 498 (1999)) or instead
of -0.003 (with Wien97, Komelj, PRB 66, 140407, 2002). I tried with other
values of RINDEX but the results are worse (and I think that the magnetic
dipole term Tz does not depend on a radial function, doesn't it?)
I think I have used the correct succession of commands to obtain this too high
value of Tz:

runsp_lapw -p -so -i 100 -cc 0.00001 -ec 0.0000001
x lapw0
x lapw1  -up
x lapw1  -dn
x lapwso -up
x lapw2 -c -up -so
x lapw2 -c -dn -so
x lapwdm -up  -so -c
x lcore -up
x lcore -dn
x mixer

and I have checked that this succession of commands (with RINDEX = 1 and
LSINDEX = 3) give values for the spin and orbital moments of hcp Co close to
litterature.

The  file Co.indmc which I used is the following:

-10.                      Emin cutoff energy
  1                       number of atoms for which density matrix is
calculated
  1  1  2      index of 1st atom, number of L's, L1
  1 8           r-index, (l,s)index

Do you have an idea why the value of Tz which I obtain is so bad, and how I
can process to make a correct calculation?

Thank you for your help,

Florian Gimbert