[Wannier] Wannier90 calculation for a1g or egPi orbitals with local coordianates and band matching

[Elias Assmann]

On 07/16/2015 03:01 PM, Santu Baidya wrote:
> I am using wannier90 code with wien2k for calculating wannier 
> orbitals corresponding to four bands crossing Fermi level.
...
> I would like to know if wannier90 capable of showing a1g or egPi
> type orbitals combination of dxy, dxz, dyz orbitals.

If you are using Wien2k, actually wien2wannier is responsible for the
initial projections rather than Wannier90, so this is really a
wien2wannier question.

The ‘write_inwf’ utility lets you choose from a number of “special”
combinations of spherical harmonics, but in the ‘case.inwf’ file, you
can make any combination of Ylm's you like [caveat: in the usual basis,
not the real-valued one].

> My second concern is band matching. For simple structure like SrVO3
> band matching is reasonable. But for complex structure even where
> four bands are separated from other bands wannier90 bands do not
> match reasonably with total band. Please tell me what should I do in
> this case.

I assume you mean the match between Wien2k bands and the Wannier
interpolated bands from Wannier90.  What you say sounds like you are
able to make a Wannier projection without disentanglement [isolated
bands].  I can say nothing beyond the usual advice:

 * Make sure that the bands are properly identified and that you really
do not need disentanglement

 * Choose proper initial projections

 * Check that the Wannierization was converged

 * Try more k-points

> Is it possible to give local coordiantes for band plotting using
> wannier90?

I do not know what that means.

-- 
Elias Assmann

Wien2Wannier: maximally localized Wannier functions
              from linearized augmented plane waves

<http://www.ifp.tuwien.ac.at/forschung/arbeitsgruppen/cms/software-download/wien2wannier/>