for the spectral function the self-energy is taken as diagonal in nn'. Giustino's El-Ph paper notes that
[...] we are considering for simplicity non-degenerate electronic states, and making the assumption that the off-diagonal elements of the self-energy can be neglected. In more general situations the right-hand side of Eq. (163) needs to be to diagonalized, or alternatively the off-diagonal terms need to be treated perturbatively.
I was browsing through the code and I was wondering if EPW checks for the magnitude of the off-diagonal elements at some point (as I would be interested in see if in some cases these are large or small)?
specfun has in the header
!! We take the trace of the spectral function to simulate the photoemission
!! intensity. I do not consider the c-axis average for the time being.
!! The main approximation is constant dipole matrix element and diagonal
!! selfenergy. The diagonality can be checked numerically.
But I do not see where this is done (I blame my very rusty Fortran)...
If someone could point to me how/where this is (or could be) evaluated I would be very grateful!
Thank you all in advance for your support!
Chris