# SIESTA error: must be in PAO.Basis, (it has semicore states)

I keep getting this error with Siesta every time I do a calculation,

Program aborted. Backtrace:
Semicore shell(s) with -1 electrons included in the valence for
Al
Semicore shell(s) with  1 electrons included in the valence for
As
Al                   must be in PAO.Basis (it has semicore states)


I'm pretty sure I'm including them properly, I have a structure with Al As and C, the C works fine, but I'm unsure why the others aren't working well.

I've tried using B instead of Al and I get the same error.

Pseudopotentials are the ones from here

https://departments.icmab.es/leem/SIESTA_MATERIAL/Databases/Pseudopotentials/periodictable-gga-abinit.html

• Can you share your full input file. It might be that you have defined the same thing twice in two different ways. It might be useful to check this thread : Dec 21, 2021 at 14:45
• Lots of views for a non-HNQ! Sep 7 at 13:17

Before I try and answer your specific query, the first thing I would do is to get my Pseudopotentials from :

http://nninc.cnf.cornell.edu/

The Pseudos here are utilized a bit more and better tested. It is also what SIESTA developers recommend here :

https://departments.icmab.es/leem/siesta/Pseudopotentials/index.html

Testing out your Pseudos should be an important step. Of course let's try and make the calculations work first !

Now, for your specific question, it would help if you could share your input (.fdf) file if possible since I think what is happening is that a Pseudo is either not well defined or that there is contradictory information in the basis set versus the Pseudos. Here is a thread of a similar issue in the SIESTA mail archive : https://www.mail-archive.com/siesta-l@uam.es/msg03437.html