Calculating electron-phonon coupling with finite difference method

[maocean]

Hi,

Recently I'm trying to calculate the electron-phonon coupling matrix with finite difference method for some special q point in the BZ, such as the Gamma point, in wannier basis. In the attachment I append the formulas I use for calculations. Specifically, the phonon frequency and the corresponding eigenvectors are obtained through the ph.x code. The wannier TB for the pristine structure and phonon distorted lattice structure are obtained through the wannier90 package. I benchmarked my method with the B-doped diamond, an example provided by EPW. The calculated |g_mn(kq)| at k=[0 0 0] and q=[0 0 0] is 775 meV while EPW gives 374 meV and the phonon linewidth is calculated to be 16.2 meV while EPW gives 4.62 meV (using exactly the same settings for the kmesh, broadening, temperature) for the three degenerate phonon modes. Could anyone help figure out the origin of the difference? Any comments and suggestions are welcome, thanks.

Haiyang
Shanghaitech University

[sponce]

Dear Haiyang,

Since you are here comparing the k=[0 0 0] and q=[0 0 0], I would first print the electron-phonon matrix element directly from the ph.x code.

Note that this can now be done using the Quantum Espresso develop version - https://gitlab.com/QEF/q-e/-/merge_requests/2116

If this is working, you can then confirm that you get the same results with EPW.

Best wishes,
Samuel