For B-doped diamond (PRB76,165108,, fsthick is ~1.36eV whereas for Pb, it is 6eV. This is probably a simple question but how should one decide on how many bands to include in the self-energy delta function? Is this a parameter that needs to be converged?

Thank you,

Vahid

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

Thu Apr 14, 2016 7:03 pm

giustino

Joined: Thu Jan 14, 2016 3:04 pm Posts: 10

Re: Criterion for choosing fsthick

Hi Vahid,

fsthick is not a convergence parameter. It can be used to speed up calculations of the electron-phonon coupling strength lambda, the phonon self-energy/linewidths, and the Eliashberg gap. The rationale for using it is that these quantities involve Dirac deltas centered at the Fermi energy, which in the code are replaced by the smearing function specified by ngaussw. This smearing function is zero when the electron energy is 5-10 times degaussw from the Fermi energy, therefore the Dirac delta will be vanishing for electron bands outside of this range. If you specify fsthick then all bands outside of this window are simply skipped in the summations (instead of evaluating a vanishing Dirac delta).

Bottom line: the results of you calculations should be independent of fsthick, what changes is the calculation time. A reasonable value is between 1 and 2 eV.

Warning: do not set fsthick smaller than 1 eV, as you may end up skipping bands which yield nonzero contributions. If you are unsure, just leave it to the default value (ie practically infinity)

Users browsing this forum: No registered users and 2 guests

You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum