Skip to content

k-ido/cif2qewan

 
 

Repository files navigation

cif2qewan

cif2qewan.py is a simple python script to create quantum-ESPRESSO and wannier90 inputs from cif files.

Usage

  1. Prepare cif2cell. (See for details, https://sourceforge.net/projects/cif2cell/)

  2. Prepare pseudopotentials in PSLibrary.

  3. Download or clone the github repository, e.g.

    % git clone https://github.com/wannier-utils-dev/cif2qewan

  4. Edit cif2cell_path and pseudo_dir in cif2qewan.py.

  5. Run.

    % python cif2qewan.py **.cif

    % pw.x < scf.in > scf.out

    % pw.x < nscf.in > nscf.out

    % wannier90.x -pp pwscf

    % pw2wannier90.x < pw2wan.in

  6. Edit dis_froz_max in pwscf.win. Recommended value is around EF+1eV ~ EF+3eV.

  7. Wannierize.

    % wannier90.x pwscf

Compare band structures of DFT and wannier90

cif2qewan.py prepares band calculation input files in directory "band".

 % cd band

 % pw.x < ../scf.in > scf.out

 % pw.x < nscf.in > nscf.out

 % bands.x < band.in > band.out

 % cd ..

 % python band_comp.py

Then, you can get the band structure plot of DFT and wannier90.

Compare band energy of DFT and wannier90

cif2qewan.py prepares nscf input file for energy diffierence. Wannier90 Hamiltonian should reproduce the band energy on the kmesh for wannierzation. (For example, 8x8x8 mesh including gamma point (8 8 8 0 0 0 in QE expression).) Here, the code checks the energy difference of DFT and wannier90 on the shifted kmesh. (c.f., 8 8 8 1 1 1 in QE expression))

% cd check_wannier

% pw.x < ../scf.in > scf.out

% pw.x < nscf.in > nscf.out

% cd ..

% python wannier_conv.py

% cat check_wannier/CONV

wannier_conv.py calculates the energy differences and outputs the result in check_wannier/CONV. average diff means \delta defined by

\delta^2 = \frac{1}{N} \sum_{n,k} (e_{n,k}^{DFT} - e_{n,k}^{Wannier})^2.

Reference

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 96.2%
  • Shell 3.8%