Posted by Steven Vancoillie on May 05, 2011 at 16:54:44:
In Reply to: CASSCF bug? posted by Tanya on April 28, 2011 at 15:20:37:
: I am trying to get the correct CAS (12/12) for a big dichromium complex and the first orbital that enters my CAS is one of the Cr 3p orbitals! which has an energy of -2 eV and clearly should be very inactive. The other 11 orbitals are fine (3d and 4s). This 3p orbital cannot be replaced by other orbital by using the ALTER command.
: I wonder if this is a bug... it happens using molcas 7.4 and 7.7.
This is not a bug, but something that is commonly encountered with transition metals. There is no single solution, you could try either (i) switching the problematic orbital with the first orbital and then freezing it, (ii) freezing the entire core, (iii) getting rid of the valence orbital that has a too high occupation and is being replaced by the core orbital, (iv) including the core orbitals (3s/3p) in the active space, (v) calculating an average of ground- and excited states such that all your active orbitals are singly occupied in at least 1 state.