Posted by Leon Freitag on August 23, 2012 at 14:15:34:
Hello,
I am trying to run a CASSCF calculation on a metal complex with a D3 symmetry. Since D3 point group is not supported in Molcas, I run in C1 symmetry and use the Supsym keyword to disable certain orbital rotations. I also use the TypeIndex keyword to define my active space. However after the calculation has finished, I have noticed that the supersymmetry vector is not correct, i.e. orbitals which should not have been rotated into each other were rotated and vice versa.
My understanding of the problem is the following: if one selects the active space by modifying the TypeIndex, at the beginning the calculation, the orbitals are rotated so that the inactive-active-secondary orbital space structure is preserved, and active orbitals are kept altogether. (just like if one uses the Alter keyword) My question is now, is the supersymmetry vector accounting for these initial rotations, or does it remain unchanged and therefore the supersymmetry is broken when the orbitals are rotated in such way?
Should I therefore avoid the TypeIndex keyword if using supersymmetry and stick to Alter keywords, adapting the supersymmetry vector accordingly?
I hope I've made my points clear. Thanks in advance to those who reply.
Leon