Mod:Hunt Research Group:aim pseudopotentials

AimAll with Pseudopotentials

From G09 B.01 AimAll should be able to handle PPs

From the AimAll website http://aim.tkgristmill.com/faqs.html#ecps

"Gaussian 09 B.01 and later support writing .wfx files and automatically writes <Number of Core Electrons> and <Additional Electron Density Function (EDF)> data sections to the .wfx file when the wavefunction is based on an ECP calculation."

This web-page has additional information as well, read it, and the following http://aim.tkgristmill.com/wfxformat.html#edfsforecps

Generating the .wfx file

Setup a standard input with the geometry in cartesian coordinates (geom=cartesian) Include the following keywords " density=current pop=no output=wfx". To the very bottom of the input file add FILENAME.wfx. FILENAME needs to be the same as the input file name, only with .wfx.

Errors we have encountered

• No wfx file generated and the following error

Warning:  atom     1 atomic number  58 has  11 electrons.
           using simpleminded designation of core shells.
          I=    1 IA= 58 I1=    1 NAtG=29 NE=  47.000 C=  52.800 Err= 5.80D+00
 StAtR1:  DoIntr=F IAtCor=2 NETot=      62 Failed=T Sum=  5.80D+00 SumC=  0.00D+00
 Final check failed in StAtR1.
 Error termination via Lnk1e in /apps/gaussian/g09_d01/g09/l9999.exe at Thu Jun 28 15:28:28 2018.

my guess is that there is a miss-match of electron density from the atomic core and the PP (this is only a guess!)

from the website above "The .wfx files written by Gaussian B.01 and later for ECP-based wavefunctions automatically include, as needed, atomic core electron densities derived from all-electron, relativistic calculations of free atoms."

this means we should look at finding a better core description

the solution is to use a small core PP with explicit 4f electrons, using the Stuttgart/Cologne ECP28MWB and ECP28MWB_SEG basis was successful in producing the .wfx file.