[Pw_forum] maximally localized wannier function and k points generated

[Andrea Ferretti]

Hi Zhou,

>   I am a newe pwscf user , I want to calcalate the maximally localized
> wannier function use pwscf together with Want,
> I have two question .

(note that, due to technical problems with the web site, want-2.1 is 
there only since few days ago; consider to update to this version if the 
case)

> I want to user molececalr dynamics to simulate different strcure  at
> temperature and pressure ,and calcualate
> the maximally localized wannier function of the differet structure
> and to see it's changs .my system is metal,
> I wonder if I can use pwscf and Want to achieve this aims-simulating
> different strcure and calculating  maximally localized
>  wannier function. 

yes, you can; maximally localized Wannier funcitons for metals are
computed following the framework of Souza, Marzari and Vanderbilt.

> I know If my system is not metal ,then CP method is
> encough ,but if I  use BODM molecular dynamics,then how
> I can calculate the maximally localized wannier function,this is the
> first question.

whatever method you use to compute the electronic structure of 
your system, espresso will export dft data in the same way (i.e. with 
the same format),
so you can compute WFs on top of any configuration you have computed...
see below for the technical aspects (input vars, etc)

>  Second question, I have searched  the pwsfc maillist    for calculating
> maximally localized wannier function , the first step is scf calculation
> ,second is nscf calculation, then run want calculation ,I feel puzzle the k
> points generation of three procedure ,are these is the same.all along the
> standard symmetry lines ?

WanT needs DFT data (the "electronic structure") on a regular (i.e. 
equally spaced) mesh of k-points in the *full* BZ;
where full means not reduced by symmetry (the possibility of using kpt 
symmetrization is under development at the moment, but not yet available 
for production).

a way to do a WF calculation using pwscf is the following:

* perform a scf calculation   (whatever kpt mesh which gives you a 
                               converged charge density) 

* perform a nscf calculation  providing a regular kpt mesh over the whole 
                              Brillouin zone. Such mesh can be generated 
                              with the kgrid.x utility of WanT

* do not forget to use wf_collect = .TRUE. in the nscf calculation, 
  or to run pw_export.x after the nscf calculation to export dft data

note that *all* these calculations aim to sample the full BZ, and 
therefore there is no reference to any "standard symmetry line"


> Thanks in advance.
> **Best regards,
> ZhouDaWei
> 

hope it helps
andrea

--
Andrea Ferretti
National Research Center S3, CNR-INFM  ( http://s3.infm.it )
Dip. Fisica, Univ. di Modena e Reggio E. (Italy)
Tel: +39 059 2055301;  Fax: +39 059 374794;  Skype: andrea_ferretti
URL: http://www.nanoscience.unimo.it

Please, if possible, don't send me MS Word or PowerPoint attachments
Why? See:  http://www.gnu.org/philosophy/no-word-attachments.html