[Wien] QTL

[Doru Torumba]

Dear all,

We are trying to calculate the 4f-DOS projected on a different basis
(a2u,t1u,t2u) using the QTL program. To understand this program we took
first a simple test case, bcc-Ho. After the convergence (with SO and LDA+U)
we created the following 'case.inq' file:
SUMA        /FULL      ( j=1/2 or 1/2,-1/2 and 1/2,1/2 extra)
DOSYM       /NOSYM     ( symmetrization is/is not performed )
-3. 3.                 ( energy window)
1.08090                ( fermi energy)
1                      ( number of atoms for which the DOS are calculated)
1 3                    ( jatom  latom)

and 'case.cf1' file with the matrix corresponding to the desired f-electron
splitting:

*0. 0.  1. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.                  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.                  a2u
 0. 0.  0. 0.  0. 0.  0. 0.  -.61237244 0.  0. 0.  0.79056942 0.  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.
 -.79056942 0.  0. 0.  -.61237244 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.
*0. 0.  0. 0.  0. 0.  1. 0.  0. 0.  0. 0.  0. 0.                  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.                   t1u
 0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  1. 0.  0. 0.                  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.
 0. 0.  0. 0.  0. 0.  0. 0.  0.79056942 0.  0. 0.  0.61237244 0.  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.
*-.61237244 0.  0. 0.  0.79056942 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.    t2u
*0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.                  0. 0.  1.
0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.    a2u
 0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0.
.  -.61237244 0.  0. 0.  0.79056942 0.
 0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  -.79056942 0.  0.
.  -.61237244 0.  0. 0.  0. 0.  0. 0.  0. 0.
*0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.                  0. 0.  0.
0.  0. 0.  1. 0.  0. 0.  0. 0.  0. 0.    t1u
 0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.                  0. 0.  0.
0.  0. 0.  0. 0.  0. 0.  1. 0.  0. 0.
 0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.
0.79056942 0.  0. 0.  0.61237244 0.
*0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  0. 0.  -.61237244 0.  0. 0.
0.79056942 0.  0. 0.  0. 0.  0. 0.  0. 0.    t2u

Running the QTL program (x qtl -c -so -up/dn) we should get a 'case.qtl1'
file with 3 columns corresponding to the a2u-t1u-t2u splitting. For spin up
we could get the 3 columns only after we removed the last 3 'stars' (*) from
the 'case.cf1' file (otherwise we have 6 columns and lots of ' ******* ').
At first sight, there is no problem for spin down, we get the 3 columns in
the qtl1-file. However, if we make the sum of the three contributions
a2u+t1u+t2u and compare this result with the total 4f-DOS obtained in a
classical way (x lapw2 -c -so -up/dn) we see a clear difference for spin
down (I can also send these graphs if somebody is interested). For spin up
the results are consistent. Therefore we think there is something wrong in
reading the cf1-file.

Best regards,
Doru Torumba

____________________________________
Doru-Cristian Torumba
E-mail: Doru.Torumba at fys.kuleuven.ac.be
Unit: Afdeling Kern- en Stralingsfysica
Address: Afd. Kern- en Stralingsfysica, Celestijnenlaan 200D, B-3001 Leuven,
Belgium