Re: wrong CASPT2 energy for phenol

Posted by Zhenyu Li on August 08, 2007 at 00:00:55:

In Reply to: wrong CASPT2 energy for phenol posted by Zhenyu Li on August 04, 2007 at 01:33:47:

thanks the comment from Prof. Roos. It turns out it's the problem of
frozen orbitals. If using all occupied states as inactive, the resulted
MS-CASPT2 transition energies are
4.79
5.93
6.66
6.80
8.04
8.10
They are within 0.2 eV to experimental energies.

: I am trying to do CASPT2 calculation on phenol in gas phase. The CASSCF energies I got are similar to what I found in literature, but my CASPT2 energies are about 0.4 eV larger. It interesting that CASPT2 did improved the CASSCF result by giving the correct state ordering, but it is systematically blue shifted.

: My CASSCF excited state energies (eV):
: 4.81 7.59 8.99 9.18 7.91 8.18
: My CASPT2 excited state energies:
: 4.98 6.33 7.11 7.19 8.23 8.28
: Experimental energies:
: 4.51 5.82 6.66 6.66

: I checked the molecular orbitals and the CI coefficients from CASSCF, it seems OK. Then it is strange why the CASPT2 energies are wrong, since we don't have many input parameters to play in CASPT2.

: My input file is following (Manually MO order changing were made between CASSCF steps):

Follow Ups:

Post a Followup

Name:
E-Mail:

Subject:

if B is 1s22s22p1, what is Li?

Passfield:

Comments:
: thanks the comment from Prof. Roos. It turns out it's the problem of : frozen orbitals. If using all occupied states as inactive, the resulted : MS-CASPT2 transition energies are : 4.79 : 5.93 : 6.66 : 6.80 : 8.04 : 8.10 : They are within 0.2 eV to experimental energies. : : : I am trying to do CASPT2 calculation on phenol in gas phase. The CASSCF energies I got are similar to what I found in literature, but my CASPT2 energies are about 0.4 eV larger. It interesting that CASPT2 did improved the CASSCF result by giving the correct state ordering, but it is systematically blue shifted. : : My CASSCF excited state energies (eV): : : 4.81 7.59 8.99 9.18 7.91 8.18 : : My CASPT2 excited state energies: : : 4.98 6.33 7.11 7.19 8.23 8.28 : : Experimental energies: : : 4.51 5.82 6.66 6.66 : : I checked the molecular orbitals and the CI coefficients from CASSCF, it seems OK. Then it is strange why the CASPT2 energies are wrong, since we don't have many input parameters to play in CASPT2. : : My input file is following (Manually MO order changing were made between CASSCF steps):

[ Follow Ups ] [ Post Followup ] [ Molcas user's WWWBoard ]