CCSD(T)-F12 with AVTZ-PP basis-set

[dennis...@gmail.com]

Dear Molpro-Community,

I did a lot of calculations with CCSD(T)-F12 and the VTZ-F12 basis set and the results were always very satisfying.
Now that I have a simple Bromide-Ion under investigation, I have to choose a basis set with pseudopotentials.
I tried to figure out the explicit basis set definition, but failed, so I decided to go with the AVTZ-PP as it seems to be the most user-friendly option.
Hence, for the CCSD(T)-F12 calculation, I use the JKFIT for Density Fitting and Resolution of Identity:

gthresh,optgrad=1.d-6,twoint=1.d-14,prefac=1.d-16,energy=1.d-8
basis=AVTZ-PP
{hf
accu,14
start,atden}
ccsd(t)-f12,df_basis=avtz/jkfit,df_basis_exch=avtz/jkfit,ri_basis=avtz/jkfit

This setting works with the use of the option explicit,ignore_bad_singl=1, yet with a warning and I have the feeling that I really should not ignore this:

 ? Warning
 ? Encountered below-occupied eigenvalues in virtual block. This may indicate serious deficiencies in the orbital or JKfit-basis sets (basis_exch). Offending values will be ignored, but
 PT2 is ill-conditioned!
 ? The problem occurs in f12_shared/Mp2Singles()

Can I choose another combination of basis sets for AO, DF and RI, which is better suited for such a CCSD(T)-F12 calculation or would it be wiser to alter some CCSD(T)-F12 settings like ansatz or level of approximation?


greetings
Dennis

[dennis...@gmail.com]

SOLVED.

In order to compute the Bromide-Ion, I have defined the number of electrons in the input as;
wf,nelec=36
which somehow messes up things, when ECP are used. With the setting
wf,charge=-1
everything works nice, and the before mentioned CCSD(T)-F12 approach with JKFIT basis set for DF and RI performs well, also with the AVTZ-PP orbital basis.

[Grant Hill]

Dear Dennis,

As the ECP has replaced a number of electrons (10 in the case of bromine), you need to account for this when specifying the number of electrons in your system.

You might also like to investigate the cc-pVnZ-PP-F12 basis sets for bromine, they will work well with the cc-pVnZ-F12 on lighter atoms. Relatively recent versions of Molpro have this built into the library, otherwise you can download from http://www.grant-hill.group.shef.ac.uk/ccrepo/ and read about the basis sets: https://aip.scitation.org/doi/10.1063/1.4893989

Best wishes,

Grant

--
You received this message because you are subscribed to the Google Groups "molpro-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
To post to this group, send email to molpr...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/f4be8546-c2b8-4586-a2a2-49c8d7b2903e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[dennis...@gmail.com]

Dear Grant,

I employed your recommended basis sets in my calculations without any troubles. Thank you very much for the hint!

greetings
Dennis