freyso

If I remember correctly, Si was removed because it was pathological (PAW-corrected norm not garantueed to be positive). I do not remember which version of the PAW set this was,...

And the way I got the PAW potentials in atomicdata format (which is the intrinsic format of atompaw): I downloaded the xml files, extracted the input file for the atompaw...

` potType = "xml"; ` see https://sxrepo.mpie.de/attachments/download/68/sphinx-manual-3.0.5.pdf Section 3.3, Table at bottom of page 13

No. Mostafa Joulaian did some work in his master thesis analogous to the Delta project, but it turned out to be quite a mess in the end, in part because...

> but does it mean we should not simply download the Si file and start using it? bare sphinx can. pyiron_atomistics will need a fix, or a manual fixup of...

Better not. I am sure there was a reason why I excluded it at the time, but this is min. 3 years ago, so I don't remember the details.

I found something - I generated in 2018 a directory named JTH-1.1, and it contains a list of PAW .atomicdata files that looks suspiciously similar to the one in pyiron....

So I suspect this is the current abinit database, and the Si potential shouldn't be used.

And here is the end of output of the PAW generator when it generates the Si potential: ``` Eigenvalues of overlap operator (in the basis of projectors): 1 -4.50102841E-04 2...

If you need a Si potential, what about the GPAW ones (I think, xml format).