ABINIT Discussion Forums

Can abinit 8.10.2 or 8.10.3 be used for hybrid functional calculations? specifically B3LYP or HSE06.

I need to do a dimer calculation for two different atom types. The Hybrid page on abinit website says I can do it in two steps: (1) do a ground-state calc in GGA, (2) do a hybrid calc.

For step 1, does ground-state assume geometric-optimization? which will mainly be the interatomic distance in my dimer calculation.

Shouldn't a hybrid run require adjustment of the distance for further minimization? Should the step 2 be also an optimization (ionmov etc) or should I keep the distance same as in step 1?

Furthermore, I noticed atompaw doesn't allow generation of hybrid PAW datasets (gives a message: "unknown libxc family -- need to work harder").

Where can I find PAW datasets generated for hybrid functionals? or is there a program that can generate such datasets (looks like atompaw doesn't)?

On further inspection of the input files shared in 'Hybrids' page on website, featuring C diamond, N2, and CH calculations, I have a couple more questions:

In N2 and CH calcs specifically, it's a 3 step process (1) ionmov with GGA-PBE (2) ionmov with PBE0 (3) ionmov with HSE06. Why was PBE0 used?

A comment says, we have to do a GGA-PBE calc first, because PAW datasets are in GGA-PBE. Does that mean we only need PAW if we are not reading _WFK from a previous run (with irdwfk or getwfk)? And that if you're able to read _WFK then all the information is encapsulated in it and you don't need PAW (e.g., for later PBE0 or HSE06 calcs)? I didn't know that.

Thanks in advance