[Wannier] Wannier Digest, Vol 108, Issue 7

[Jonathan Yates]

On 6 Feb 2017, at 15:40, KuangChung Wang <kevin60105 at gmail.com> wrote:

> Hi
> 
> I have a question with the hybrid orbitals for the projection process. 
> 1. for trigonal prismatic structure, a sd5 hybrid orbital is needed. However, I checked online, with almost no info found for the coefficient to linearly combine sd5 orbitals. There are some rules like sum(row^2)=1 sum(col^2)=1 or orthogonal.  And some suggest group theory needed for the coefficient. Does anyone know the coefficient?
> https://en.wikipedia.org/wiki/Trigonal_prismatic_molecular_geometry
> 2. After finding out the coef, it should be implemented in wannier90-2.0.1/pwscf/v5.0/pw2wannier90.f90  ?  However, this seems to be working only with Quantum espresso?

pw2wannier90 is the interface to PWSCF. So if you are using PWSCF you need to modify that routine - if you are using another code then you would need to modify the corresponding interface.

 If you add new types of projection there are some changes needed to Wannier90, however, these are trivial book keeping.

 I don’t know the coefficients you need - I’d have to work them out. However, I think that you first want to prove that you do need this projection.

 In the pre-wannier90 days we wanted to find the WF for bcc iron. At that time we could only project onto s,p,d orbitals. So I did this - and found that when I left the minimiser go for a few iterations it always lowered the spread by hybridising the orbitals to a set of sp3d2 and 3 dxy orbitals. Having realised this we then implemented the projection onto the sp3d2 hybrid orbitals. Have you done something similar for your system? Do you know that the MLWF will have this sd5 symmetry? I would check before going through the effort of coding up the projections.

 Jonathan


-- 
Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK
tel: +44 (0)1865 612797                http://users.ox.ac.uk/~oums0549/