Possible bug in Sodium semi-core GGA PW91 pseudopotential: pot_GGA/Na_pv

Message

Brane · #1 Post by **Brane** » Tue Apr 01, 2008 2:24 am

Hi,

When I do some calculations using Sodium semi-core GGA PW91 ultrasoft pseudopotentials (located at /pot_GGA/Na_pv), an error was triggered:

ERROR FEXCP: supplied Exchange-correletion table
is too small, maximal index : 3922

However, when I switched the potential to PAW GGA PW91, everything is Ok. By the way, I am using the latest VASP 4.6.34.

I notice that a similar error message was also reported in a previous thread :
http://cms.mpi.univie.ac.at/vasp-forum/ ... .php?3.462

"When I doing a calculation for bcc sodium, getting POTCAR file from pot_GGA/Na_pv/ . I was informed of the error message :

ERROR FEXCP: supplied Exchange-correletion table is too small, maximal index : 3875"

I also notice that there is a tag "TITEL = NC Na" in Sodium Na_pv potential. This is strange, because this tag is "TITEL = US" for all other elements. I guess that NC stands for Norm-Conserving, while US for UltraSoft.

I'd like to mention that nothing is wrong with the geometries and setting parameters. No matter which IBROIN is used (1 or 2), this error is always there.

Based on above-mentioned facts, I suspect that there could be a bug in Sodium semi-core GGA PW91 ultrasoft pseudopotentials.

Thanks for your kind comments in advance.

Best regards,
Brane

#2 Post by **admin** » Mon Apr 07, 2008 7:59 am

1) please first of all check if the POTCAR file you use is the most up-to-date one as can be downloaded from the ftp server.
2) please always use potentials of most similar hardness