sg-15 got abnormal gw correction value.

[pangrt]

Dear all,

I calculated a system using sg-15 and dojo at the same time and got inconsistent results.

The cb+1 energy obtained by using the sg-15 pseudopotential after correction reaches below cb, resulting in a decrease in the overall gap after GW correction, which is inconsistent with the results obtained by the dojo pseudopotential (the gw result of dojo is a normal band gap increase) .

I'm not sure if there's something wrong with my calculations or if it's caused by the sg-15 pseudopotential?

Thanks!

[Daniele Varsano]

Dear RongTianPang,

the main difference between sg15 and pseudodojo is the presence of the non-linear core correction in pseudodojo set.
Yambo by default ignores these corrections, and they can be taken into account using the keyword UseNLCC in the input file.
If it is better to include or neglect them is not totally clear, as they enter the Vxc only:
You can have a look at this thread (viewtopic.php?p=13451&hilit=UseNLCC#p13451)

As first check anyway, it is recommended to look if the two pseudos contain the same number of valence electrons (semi-core or not).

Best,
Daniele