[Wannier] Refs for constructing M and A matrices

[Jonathan Yates]

On 13 Nov 2017, at 01:17, Hung Pham <phamx494 at umn.edu> wrote:

> I am looking for a detailed reference (rather than PR B 1997 and PR B 2001) that shows how to construct the Mmnkb and Amnk matrices using the u_nk from an electronic structure code and ouput (l, mr, r, etc) of  wannier_setup. I am working on making an interface between an electronic code and wannier90.
> 
> I was trying to look at pw2wannier90 source code to learn from it but there are many variables and functions that belong to QE that I am not familiar with. So I need a practical reference.

The Wannier90 user guide is very clear on what the Mmn and Amn files should contain. What it cannot do is tell you how to compute these in your specific code - why? because every code is different.

 To make some general comments:
    The Amn is usually pretty simple - as most codes already have the ability to project onto local orbitals of some form (e.g. for population analysis)
  The Mmn is trickier as you require overlaps between wave functions at different k-points. This operation doesn’t usually appear as part of, say, calculation of the ground state energy. But if your code has routines to compute Berry Phase / polarisation then you will find these overlaps and can probably re-purpose the routines.

 If you tell us more about the code you are hoping to interface we might be able to make more specific suggestions. In any case the best people to ask for advice will be the developers of the code you are using.

Jonathan


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