Why do the PBE pseudopotentials for Zn and S from pseudo-dojo give such bad lattice constant (~10%) in siesta but it's reasonable for the PBEsol(~1%) or even LDA pseudopotentials(~3%)?

How to generate my own pseudopotential in psml format for siesta 5.2?

Delta SCF for periodic system

Following up on @Schoblerone's comment in this answer to How to obtain a defect configuration coordinate diagram using first-principles calculation?:

After some reading, I think Δ SCF procedure doesn't work for periodic system. I tried manually fixing the occupation to simulate an excited (promoted 1 electron in the highest occupied defect level to a higher level) state but self-consistency was not reached. Is it possible to obtain this within the realm of DFT by maybe making the system charged and then relaxing the coordinates?

I'd very interested how you came to that conclusion. Regarding your calculation: ΔSCF often does not work out of the box as easily as regular DFT calculations.Can you describe your system further? Especially relevant: are you exciting into a degenerate level? Unless you have a solid case to make that the defect is excited by some kind of charge switching, I don't believe adding an electron is a good idea.

References that says delta SCF is not possible for periodic system:

Explanation on why it should not work:

References where they have used delta SCF for periodic system and it worked:

Question: Why is the discrepancy

Why do the PBE pseudopotentials for Zn and S from pseudo-dojo give such bad lattice constant (~10%) in siesta but it's reasonable for the PBEsol(~1%) or even LDA pseudopotentials(~3%)?

How to generate my own pseudopotential in psml format for siesta 5.2?