real axis GW calculation

[samaneh]

Hi there,
I run yambo -g n with the optimized parameters (e.g. EXX, NGs)
of running yambo with plasmon pole approximation,
but the band gap energy is larger than the
calculated one with yambo -g n -p p!!
Would you please explain if there is a relation between the optimized
parameters needed for these two commands(yambo -g n and -g n -p p)??
Thanks
Best
S. Ataei, university of Tehran, Iran

[Daniele Varsano]

Dear Atei,
the real axis integration approach (yambo -d -g n) do integrate the self energy (GW) in the energy range given by GbndRnge and an important parameter to converge is the energy step used for the integration (ETStpsXd). This is hard to converge, the dynamical screening is calculated for all the energy points and for a large number of points the calculations turn to be very cumbersome. The PP is a model, and rely on the calculation of the screening in two energy points (zero and the imaginary frequency (PAPntXp). Please note that the two methods are not equivalent, the PPA is an approximation which assume that the frequency dependence of the screening has a single pole.

Best,
Daniele

[samaneh]

dear Daniele,
Thanks
I know the difference between PPA and real axis calculations
but I asked you about the convergence tests and parameters optimization
you mean along with doing convergence tests for EXXRLvcs and NGsBlkXd and GbndRnge
(which are used for PPA model) I should perform convergence tests for ETStpsXd for real axis calculations?
Best
S. Ataei

[Daniele Varsano]

Yes,
and this is the problematic issue.
Daniele