Molcas Forum

Mohammed Obies

Member

Registered: 2017-08-11

Posts: 5

RASSCF ciroot and CASPT2

Hi

I have calculated all roots for specific symmetry and I need to calculate the CASPT2 energy for all roots. For example, here I have 4 roots and I need to calculate CASPT2 energy for all 4 roots. However If I use just CASPT2 keyword it will be calculated the energy for just first root. How can calculate the energy for all roots?

Natural orbitals and occupation numbers for root  1
      sym 1:   0.999414   2.000000
      sym 2:   2.000000
      sym 3:   2.000000
      sym 4:   1.000587   2.000000

      Natural orbitals and occupation numbers for root  2
      sym 1:   1.501037   1.504109
      sym 2:   2.000000
      sym 3:   2.000000
      sym 4:   1.496113   1.498741

      Natural orbitals and occupation numbers for root  3
      sym 1:   2.000000   1.999999
      sym 2:   0.999213
      sym 3:   1.000789
      sym 4:   2.000000   1.999999

      Natural orbitals and occupation numbers for root  4
      sym 1:   2.000000   1.011097
      sym 2:   1.999999
      sym 3:   2.000000
      sym 4:   2.000000   0.988905

Thanks

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,202

Re: RASSCF ciroot and CASPT2

The easiest is:

&CASPT2
* four roots, roots number 1, 2, 3 and 4
Multistate = 4  1 2 3 4
* but skip the MS part, unless you are interested in an MS-CASPT2 calculation
NoMultistate

Mohammed Obies

Member

Registered: 2017-08-11

Posts: 5

Re: RASSCF ciroot and CASPT2

Thank you for your answer. I have done that I get the energy for CASPT2 but the energy of root1 equal to -507862.0963076590eV and for root4 equal to -507869.9285553810eV and it seems the root4 has lower energy than root1, do you think that reasonable? because I think the energy of root1 should has the lowest energy of all roots.

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,202

Re: RASSCF ciroot and CASPT2

Root order can change between CASSCF and CASPT2, that's common and normal. It means, of course, that you should be careful with the interpretations: the CASPT2 results could be affected by a larger error (due to poor CASSCF reference), the selected CASSCF states could be not the relevant ones for CASPT2, etc.