ABINIT Discussion Forums

Dear memebers of this community: Although I am newbie in using ABINIT I though I understood the basic features of it until I confront myself to the example of positron calculations (/abinitx.x.x/tests/v4/Input/t95.in and /abinitx.x.x/tests/v4/Input/t96.in ). I found that for running that test (seeing the /abinitx.x.x/tests/v4/Refs/t95.out and /abinitx.x.x/tests/v4/Refs/t96.out ), the computation of course needed a pseudopotencial file for the atoms used (in these case Aluminium) and the one used this time was 13al.psppos file. Reading also the stuff from internet I found that the pseudopotencial files used for positron calculations must be the FHI (Fritz-Haber-Institute) pseudopotentials
and indeed the file from 13-Al.LDA.fhi (the one that it is open if you click on Aluminium on the web page of FHI pseudopotencials, http://www.abinit.org/downloads/psp-lin ... ks/lda_fhi) is very similar to 13al.psppos one. Similar but not equal! in fact, the files looks exactly the same but 13al.psppos has a block at the end that 13-Al.LDA-fhi does not!! Trying myself to make positron calculus for other elements I notice that I could not do it properly with the pseudopotencial files .fhi; meaning that this difference between 13-Al.LDA.fhi and 13al.psppos file is important to make this calculus right. The problem then resides on that: How to "convert" from .fhi file to a one .psppos???
Any ideas???

I see that the post now has already a couple of views but no answers. I beg anyone that visits me leave me a comment, independently that they know how to solve this problem I am mentioned. I am interested if my post is clear or if you visiters have any question to me.
Regards,
Sergio

dr_sferrari wrote:Dear memebers of this community: Although I am newbie in using ABINIT I though I understood the basic features of it until I confront myself to the example of positron calculations (/abinitx.x.x/tests/v4/Input/t95.in and /abinitx.x.x/tests/v4/Input/t96.in ). I found that for running that test (seeing the /abinitx.x.x/tests/v4/Refs/t95.out and /abinitx.x.x/tests/v4/Refs/t96.out ), the computation of course needed a pseudopotencial file for the atoms used (in these case Aluminium) and the one used this time was 13al.psppos file. Reading also the stuff from internet I found that the pseudopotencial files used for positron calculations must be the FHI (Fritz-Haber-Institute) pseudopotentials
and indeed the file from 13-Al.LDA.fhi (the one that it is open if you click on Aluminium on the web page of FHI pseudopotencials, http://www.abinit.org/downloads/psp-lin ... ks/lda_fhi) is very similar to 13al.psppos one.

this is just the format chosen, doesn't mean it is generated with fhi98pp

Similar but not equal! in fact, the files looks exactly the same but 13al.psppos has a block at the end that 13-Al.LDA-fhi does not!! Trying myself to make positron calculus for other elements I notice that I could not do it properly with the pseudopotencial files .fhi; meaning that this difference between 13-Al.LDA.fhi and 13al.psppos file is important to make this calculus right. The problem then resides on that: How to "convert" from .fhi file to a one .psppos???
Any ideas???

The final block is (probably) the non-linear core correction. I don't know the details of the positron pseudopotential. You should check the literature, and see if the contributors from CEA have a reference or 2. Read CPC 180 1582 Gonze et al. in particular, which should have further in depth references. Also check if this has been announced as a production feature (in the release notes for 5.9 and 6.0) to make sure that it _should_ be working properly in the general case.

Matthieu