KRMCSCF in linear geometry (with LUCIAREL) in Dirac 19 not working for me

[jvi...@gmail.com]

Dear Dirac experts,

I am calculating KRMCSCF for NU+ ion, which ground state shall be Omega = 4. 

I did run SCF and PBE in linear geometry and the Omega=4 state could be represented on this level by open-shell with two singly occupied U-5f-based MOs with m_j = 3/2 and m_j = 5/2.

I took the PBE orbitals as initial guess and tried MCSCF with LUCIREL and set .SYMMETRY to 8, as documentation suggested (double the Omega value, https://www.diracprogram.org/doc/release-19/manual/wave_function/krmcscf.html).

I took (6,6), which covers both the U-5f-based MOs well, with m_j = 3/2 and m_j = 5/2. But DIRAC told me in the output that there are no determinants for this symmetry :-(

Please give me a hint about where I need to correct a mistake.

The output, input+geometry and script are attached. Orbitals in DFPCMO are to big to be attached.

I am trying to avoid GOSCIP for a CI-like subroutine of MCSCF as it does not work with linear symmetry but only C2 (not C2v, right?), where Omega=4 could end up to be an excited state* within one of the two co-irreps, so I could not access it. 

Best regards,

Jakub

* As in SCF, in the tiny AO basis I am yet using (contracted DZ), the correct order of states might not be described, and Omega=4 might not be the ground.
Also, is it possible to access excited states (of given symmetry) by KRMCSCF in Dirac (perhaps some higher version), please?