Page 1 of 1

Wannier-Interpolated eigenstates

Posted: Wed Mar 29, 2017 10:17 pm
by Vahid
Dear EPW Users,

As you are aware, the initial eigenstates (u_nk), the periodic part of the Bloch states, are available from DFT wave functions in the evc.dat files from QE.

Does EPW store (or calculate) the "final" Wannier-interpolated eigenstates in the Bloch representation for the fine or coarse k-grid? I am hoping to use this information to calculate the overlap integral u*_k'(r)u_k(r)dV within EPW which can then be used for calculating impurity scattering rates.

Thank you,

Vahid

Vahid Askarpour
Department of Physics and Atmospheric Science
Dalhousie University,
Halifax, NS, Canada

Re: Wannier-Interpolated eigenstates

Posted: Wed Mar 29, 2017 10:42 pm
by carla.verdi
Dear Vahid,

While the Bloch states on the coarse grid are available from the DFT calculation, EPW doesn't calculate the Wannier-interpolated eigenstates in the Bloch representation on the fine grid - nor does Wannier90 actually. It can be done but it's expensive and generally not required.

Best
Carla

Re: Wannier-Interpolated eigenstates

Posted: Wed Apr 05, 2017 12:54 pm
by Vahid
Dear Carla,

Since calculating Bloch states on the fine grid is expensive, to calculate the overlap integral for the periodic part of the Bloch states (u_nk) on the fine grid, can one calculate the overlap integral on the coarse grid (as done in the Wannier90 code by calculating M_mn matrices) and then interpolate M unto the fine grid using the same procedure done for the Hamiltonian and el-ph matrix elements?

Thank you,

Vahid

Re: Wannier-Interpolated eigenstates

Posted: Wed Apr 05, 2017 1:58 pm
by carla.verdi
Dear Vahid,

It can probably be done but it's not trivial if you try and formally invert the formulas in Bloch/Wannier representation for the overlaps of the Bloch states.

Carla