Molcas Forum

chemphys

Member

Registered: 2020-11-10

Posts: 27

[SOLVED] Mixing ECP and basis set in seward

Hi,
For seward input files I've been using the format:

 &SEWARD &END
Basis Set
Sr.DEF2-SVP
Sr2       2.32009000    3.03510900    -0.74537300   Angstrom
End of Basis
Basis Set
Cl.DEF2-SVP
Cl3       -0.23276700    0.55816800    4.27323100   Angstrom
End of Basis

If I wanted Sr to have a different ECP than what corresponds to DEF2-SVP (but is still listed in the basis_library), how would I specify that in this format?  Is it possible to do this in one line?  Such as:

 &SEWARD &END
Basis Set
Sr.ECP.X / DEF2-SVP?
Fe2       2.32009000    3.03510900    -0.74537300   Angstrom
End of Basis

or something similar?  Thank you!

Last edited by chemphys (2021-02-22 21:13:21)

jxzou

Member

Registered: 2020-09-11

Posts: 14

Re: [SOLVED] Mixing ECP and basis set in seward

Yes, you can use mixed ECP and basis sets as you desired. The following 3 steps would take you several seconds to accomplish this:

(1) write/create a Gaussian .gjf file like

%chk=Sr_mix_ecp.chk
#p UHF genecp nosymm int(nobasistransform) guess(only,save)

title

0 2
Sr       2.32009000    3.03510900   -0.74537300
Cl      -0.23276700    0.55816800    4.27323100

Sr 0
def2svp
****
Cl 0
cc-pVDZ
****

Sr 0
LANL2
[blank line]
[blank line]

(2) Submit it to Gaussian. This takes <10 seconds since the keyword 'only' is specified. Then you can run 'formchk Sr_mix_ecp.chk' to generate a .fchk file.

(3) Run 'fch2inporb Sr_mix_ecp.fchk -uhf' to get the desired OpenMolcas .input file, in which basis sets data and ECP are well written. If it is a RHF-type wavefucntion, the argument '-uhf' is not needed. A .INPORB file will also be generated, but in this case it is useless since no SCF computation is done in Gaussian.

The utility fch2inporb can be found at https://gitlab.com/jxzou/mokit. In this way, you can in principle use any basis set mixed with any ECP.
Good luck.