“…Initially, all valence π-type orbitals were included in the active spaces of the analyzed molecules using the state-averaged SA(7)-MCSCF(1+2) method, which was restricted to single and double excitations within an active space with respect to a reference closed-shell configuration. We used the following active spaces: (14,13), i.e., 14 electrons distributed over 13 orbitals, for compound 1, (20,19) for compound 1A, (28,27) for compound 2A, (36,35) for compound 3A, (28,27) for compound 4A, (18,17) for compound 1B, (22,21) for compound 2B, (30,29) for compound 3B, (18,17) for compound 1C, (22,21) for compound 2C, (18,17) for compound 1D, and (22,21) for compound 2D. The active spaces were then reduced for the subsequent complete active space self-consistent field calculations at the SA(7)-CASSCF (14,14) level of theory.…”