[Wien] About isplit!!

[Saeid Jalali]

Hi Zu,
The ISPLIT comes back to its proper initial value since it is set by
symmetry program to specify the decomposition of lm. This is done based on
the symmetry of your case. Thus to be not meaningless changing ISPLIT,
proper knowledge about the new symmetry, i.e. lattice type or more
specifically space group, atomic positions, multiplicities/equivalency or
more specifically point groups, is also required. In this case, one needs
only once (with no need to go through a complete SCF) to rerun lapw1 and
lapw2: x lapw1 and x lapw2 -qtl.

Sincerely yours,
Saeid.

>   I am trying to plot the DOS for a system including the transitional Co
element with space
> group Pnnm. The case.struct shows the parameter ISPLIT=8, in order to get
the t2g and eg
> partial DOS instead of the dxy,dyz,dxz,dx2y2,and dz2 DOS for Co d states,
I forced the
> ISPLIT to 2, however I found that the ISPLIT is still the initial value
(8) after running
> the SCF. If any one can give me an answer?