Checking VXCRLvcs convergence

[milesj]

Hello,

In the Hartree Fock tutorial, it recommends checking E_xc in the yambo output against a corresponding value in the quantum espresso output to confirm VXCRLvcs is large enough. The only possible corresponding value I see is "xc contribution", which is printed at the end of the scf run. For all of my runs, these values don't seem to match - the quantum espresso output is about -265 Ry, whereas in yambo it is about -196 Ry.

I also decided to check if these values match for a few other materials: Silicon, hBN, and MoS2. For the input files and pseudopotentials I used, for some reason the values matched for MoS2, but were very different for Silicon and hBN. What could be causing this difference? It doesn't seem like a convergence issue - for my material when VXCRLvcs goes from 200000 to 400000, the value of E_xc stays around -196 Ry.

I've attached my quantum espresso input and output for hBN, as well as the yambo output.

Best,
Miles

[Daniele Varsano]

Dear Miles,

a possibility here is that you are using pseudopotentials containing non-linear core corretions.
In this case you need to activate their use in Yambo uncommenting the flag "UseNLCC".
Best,

Daniele