real axis calculation

[samaneh]

Dear developers,
I have calculated the band gap of TiO2 within PPA (yambo -g n -p p)
which is 3.6 eV (the experimental value is 3.2 eV)
but the one calculated with full frequency calculations is around 3.8 eV
(with the same parameter used for PPA (k-points, nbndrnge, EXX) but the ETStpsXd is different)
I run these calculations with 3 different ETStpsXd
1-ETStpsXd=100(default one as in the PPA calculations) bandgap=3.84 eV
2-ETStpsXd=200 bandgap=3.80 eV
3-ETStpsXd=600 bandgap=3.82 eV

As you see although I increased the value of ETStpsXd parameter
but the band gap does not decreased !!!
would you please explain me about this problem.
Thanks
S. Ataei, University of Tehran, Iran

[Daniele Varsano]

Dear S. Atei,
why the PPa approximation and the real axis integration should give exactly the same results? I cant' see the problem.
They are two different implementation, the PPa is an approximation. Next note that in PPA approximation ETStpsXd does not have any sense, it is a single pole approximation, and the model needs only 2 frequencies to fit the dielectric matrix, please have a look to the documentation, so that variable it is not read by the code when calculating the screening in PPA.
Finally note that also the Blocks in G-space (NGsBlkXp) and the bands for the screening (BndsRnXp) need to be converged.
Best,
Daniele