how to select hgh pseudopotentials?

[marcindulak]

Hi,

i would like to perform PBE calculations on H2O molecule,
with HGH pseudopotentials, atk-12.2.r1.622604e-linux64.bin.

I have three questions:
1.
Guessing from # http://quantumwise.com/forum/index.php?topic=1354.0 i would set:
basis_set = [
    GGABasis.Hydrogen_DoubleZetaPolarized(
    pseudopotential=NormConservingPseudoPotential('normconserving/HGH/H.LDA.1.zip'))
    GGABasis.Oxygen_DoubleZetaPolarized(
    pseudopotential=NormConservingPseudoPotential('normconserving/HGH/O.LDA.6.zip')),
    ]

Is it the right way?
There is no HGH pseudopotential for hydrogen (more elements are missing) under
normconserving/HGH/, but they are provided
at http://www.abinit.org/downloads/psp-links/psp-links/hgh

2.
Moreover, I'm unable to unzip the normconserving/HGH/O.LDA.6.zip file,
in order to verify if it is the expected pseudopotential:

$ ls pseudopotentials/normconserving/HGH/O.LDA.6.zip && unzip pseudopotentials/normconserving/HGH/O.LDA.6.zip
pseudopotentials/normconserving/HGH/O.LDA.6.zip
Archive:  pseudopotentials/normconserving/HGH/O.LDA.6.zip
  End-of-central-directory signature not found.  Either this file is not
  a zipfile, or it constitutes one disk of a multi-part archive.  In the
  latter case the central directory and zipfile comment will be found on
  the last disk(s) of this archive.

Is it expected?

3. Does ATK offer only basis set representation? I'm interested in getting
   an accurate HGH result, in order to compare with other codes
   performing HGH, like abinit.

[Nordland]

1) The basis set / pseudo-potential you are trying use, is the preview version from ATK 11.8. There is included the complete set bundled with entire new basis set system. So you can now write

Code

basis_set = [BasisGGAPBEHGH.Hydrogen_1_Tier_3, BasisGGAPBEHGH.Oxygen_6_Tier_5]

This means that you will use the HGH GGA-PBE pseudopotential for Hydrogen with one electron and the associate tier 3 basis set, and HGH GGA-PBE pseudopotential for Oxygen with 6 electrons and the associate tier 5 basis set. The higher the tier the better the basis set is. This is the new right way. 2) The HGH pseudopotential has been converted to UPF (since HGH does not contain an electron density), and therefore they have been encrypted. If you want to check that it is really the original HGH pseudopotential you can download it compare it like this:

Code

BasisGGAPBEHGH.Hydrogen_1_Tier_3(NormConservingPseudoPotential('./1h1.hgh')

3) ATK-DFT "only" has localized basis sets, so if by basis set representation you mean localized basis set, then yes. There is no native planewave code in ATK-DFT. The result compare rather well with abinit. Here I have attached a small test example I did of Germanium using the same HGH pseudopotential.