I am using SIESTA 3.2, which seems to only support the .psf pseudopotential format, is there any tools that allow me to convert the pseudopotential files into another format? Or when generating the pseudopotential, could I assign another format?

I am trying to use the norm-conserving pseudopotential Ce element, but there is no norm-conserving pseudopotential of rare earth metal.
May I ask if the norm-conserving pseudopotential is not fit for rare earth metal?

Hello all,

Is it possible to generate a pseudo potential for Ru using the actual electron configuration of ⦏kr⦐ 4d⁷ 5s¹ instead of the one that is used at the moment which gives it a valence of 16.

As it is, this pseudo potential is impossible to use for the material SrRuO3 in a pseudo-cubic structure. I managed to find only one pseudo potential for Ru with valence 8 but is pbe instead of pbesol and it is ultrasoft.

I found it here: https://nninc.cnf.cornell.edu/psp_files/Ru.pbe-n-van.UPF

Thank you

Dear pseudo-dojo people,

thank you very much for your work here! I just found one problem with one of the pps from your project:

The PBE pseudopotentials for K, Ta and O seem to be fine for lattice parameter calculations, but trying the PBEsol versions, I found out that the FR pps minimize the energy at a lattice parameter of around 4.3 A, which is almost 8% off the experimental one of 3.9885 A (see attached figure showing the total energy difference versus different lattice parameters), calculated with QE 6.3.

Calculations with the SR PBEsol pps end up at 3.977 A, while both PBE pps end up around 4.02 A (see attached images, which is in reasonable agreement with the experimental one.

Do you maybe know why the FR PBEsol are off that far? Is this a known problem?

Best,
Tobias

PBEsol FR with problems

PBEsol SR

PBE FR

Dear developers,

When I try to download the UPF format pseudos for NC FR (ONCVPSP v0.4), PBEsol on pseudo-dojo.org, I get the wrong format (psp8 instead of UPF). Could you re-upload the pseudos with the correct format online? Thank you very much!

P.S. This happens only for elements Ge, Sb, Te (not sure if other elements have the same problem). If you download them, the filename extension is ".upf" (which is correct), but the content of the files are of the psp8 format instead of the upf format. Quantum Espresso cannot properly read these pseudos.

Best regards,
Jingyang

While adding support for the v0.5 PBE pseudos to aiida-pseudo, I noticed that the .djrepo files for the new pseudos in v0.5 seem incomplete (e.g. http://www.pseudo-dojo.org/pseudos/nc-sr-05_pbe_standard_djrepo.tgz). For example:

{
"basename": "Ba-sp.psp8",
"ecuts": [
21.0,
23.0,
25.0,
27.0,
29.0,
31.0,
33.0,
35.0,
37.0,
39.0,
41.0,
49.0,
51.0
],
"md5": "7658a2694ebe9008fe2b3c8e9ef16c8e",
"ppgen_hints": {
"high": {
"ecut": 31.0,
"pawecutdg": 31.0
},
"low": {
"ecut": 28.0,
"pawecutdg": 28.0
},
"normal": {
"ecut": 29.0,
"pawecutdg": 29.0
}
},
"pseudo_type": "NC",
"symbol": "Ba",
"version": "1.0",
"xc": {
"@class": "XcFunc",
"@module": "pymatgen.core.xcfunc",
"c": {
"@class": "LibxcFunc",
"@module": "pymatgen.core.libxcfunc",
"name": "GGA_C_PBE"
},
"x": {
"@class": "LibxcFunc",
"@module": "pymatgen.core.libxcfunc",
"name": "GGA_X_PBE"
}
}
}

Specifically, I'm missing the following keys:

• md5_upf/md5_psml: we use this md5 to double check that the metadata corresponds to the correct pseudo file.
• hints: am I correct that the hints didn't change versus v0.4.1?

Would it be possible to update the .djrepo files to include this information? Also happy to have a crack at this, but I might need a bit of guidance. ^^

One of our users needs to convert from psp8 to UPF and since pseudo_dojo has files in both formats, I wonder if these were created manually or by using an automated tool to convert between the two. If you know the answer, could you suggest something for this user on MMSE?

On a Q&A + conversation on the Matter Modeling StackExchange site there are reports of the non-hovered options of the dropdown menus having white color fonts (on white background), so users cannot read the options unless they hover over them (see the first snapshot in the URL above).

Dear pseudo-dojo developers,

I noticed that for oxygen, the pseudo for LDA is missing on the website. Although, I found it in the GitHub repo here: https://github.com/abinit/pseudo_dojo/tree/master/pseudo_dojo/pseudos/ONCVPSP-PW-PDv0.4/O

Is there a reason why it is not on the website?

Many thanks

The LDA (Perdew-Wang) PSP for copper called cu-sp-high.in shows large deviation in log derivatives for d- and f-states. I guess the f-state is less serious but is there a reason the d-state shows large discrepancy? Interestingly, trying to rectify the situation results in negative phonon frequencies for many q-points in the BZ. In its present state, all frequencies are positive.

Thank you,
Vahid

Hi, would it be possible to generate files in .recpot format, native to CASTEP and ONETEP? OPIUM does recpot, but Rappe's group library has much smaller element coverage...

Consider pulling the http://www.thch.uni-bonn.de/tc/downloads/GMTKN/GMTKN30/GMTKN30main.html data available in https://wiki.fysik.dtu.dk/ase/.
Basic functionality is illustrated below, feel free to convert it the way you want:.

cd /tmp
svn co https://svn.fysik.dtu.dk/projects/ase/trunk ase
cd ase
cd ase/data
PYTHONPATH=../../:$PYTHONPATH PATH=../../tools:$PATH python gmtkn30.py
cd -
cat <<'END'> gmtkn30_test.py
from ase.data.gmtkn30_g2rc import data
from ase.data.gmtkn30_g2rc import info
from ase.structure import molecule
from ase.calculators.emt import EMT

# reaction number N in the g2rc set
N = 1
energy = 0.0
for compound, stoichiometry in info['reactions'][N][:-1]:
    e = molecule(compound, data=data, calculator=EMT()).get_potential_energy()
    energy = energy + e * stoichiometry
print energy, info['reaction energy']['reference'][N]
END
PYTHONPATH=.:$PYTHONPATH PATH=tools:$PATH python gmtkn30_test.py

Since carbon is kind of an important element (ask an organic chemist), it is somewhat confusing that is missing from the LDA set.

Is there a fundamental problem? As a workaround, is it possible to use the v0.3 file, or does it have some issues?

Thanks!

Link in project header appears broken (11 July 2017) http://www.pseudo-dojo.org/dojo.html