I am considering calculating the electron-phonon coupling in some correlated materials, which requires a Hubbard U. I wonder whether Abinit has this functionality.
A future version of PERTURBO may have this functionality (as indicated in the attached screenshot), but not in the current version.
The current version of QE and EPW can calculate phonons for +U systems but cannot calculate the electron-phonon coupling.
So I wonder whether ABINIT could calculate the electron-phonon coupling for +U systems.
This is possible in abinit, but only with the legacy electron phonon code. You run dfpt with PAW+U and get GKK files as prescribed in the legacy tutorial, then the anaddb code doesn’t care where they came from. Means you can do eliashberg functions and linewidth, and Allen style transport for metals. The kpt convergence has to be done by increasing the nscf density in the gkk calculations, and the qpt is interpolated using a FT which is not as efficient as the newer algorithms: you need to ensure both are well converged
Note that this has not been used very widely so there may be bugs.
We are working on porting the new eph code to PAW, which will then work with +U as well.
