[Wannier] Data grids in XSF files generated by Wannier90

[Jonathan Yates]

On 25 Feb 2010, at 14:35, Rok Zitko wrote:

> Hi,
>
> For reference, I'm posting this discussion of the format of XSF  
> files as
> generated by Wannier90.
>
> Regards,
> Rok
>
>
>> Dear Arash,
>>
>> I think I understand: rather than writing the full "general grids",
>> as expected in xcrysden, the code actually outputs "periodic grids"
>> (i.e., it omits the redundant periodic replicas) and compensates for
>> the difference by subtracting 1, i.e. rescales by (ngx-1)/ngx or,
>> generally, (ngs*ngx-1)/ngx. So the output files strictly speaking do
>> not
>> conform to the XSF specification, but they are still correct
>> as long as a single such cell is used. Of course, there is no need
>> to ever increase "Number of units shown" from 1 in xcrysden, so there
>> is no real problem in practice.

Rok,

  I agree that this is confusing topic - I keep meaning to add extra  
documentation into the code...

But I disagree with some of what you say above.

Firstly - proof by demonstration:
   It is often very helpful to increase 'number of units shown' - this  
means that you can see a periodic array of wannier functions. This can  
help to understand how they are orientated. When you do this you will  
see that Xcrysden puts them in the correct position (it is very easy  
to spot if there is an error in the xsf file!). So given it appears to  
do the correct thing we have to understand why.


First, note that the WF do not live in a space with the periodicity of  
the unit cell. Rather they live in a super-cell of the unit cell  
(which depends in the k-grid used to construct them - a 5x5x5 k- 
sampling means the WF live in a 5x5x5 supercell).

A key point is that the xsf data grid does not have to bare any  
relation to the unit cell. In fact we use this when we make a  
supercell in which to plot the WF. But the grid does not have to span  
an integer number of units cells (indeed as you have noticed for  
num_plot_supercell=1 our grid is slightly smaller than the unit cell).  
So our grid is a valid general grid.

For a periodic quantity (such as charge density) this would not be the  
correct thing to do. But for our case it is fine. One can generate all  
the WF for the system by applying translations of the unit cell  
vectors. This is what happens when we increase "Number of units  
shown" (ie it redraws the general grid shifted to the new lattice sites)


I suspect we are using xcrysden in a way that the author didn't  
originally envisage - but as far as I can tell it all works out fine.
  As Arash mentioned, streamlining the plotting routines is on our to- 
do list.

  Yours
   Jonathan



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