Dynamic correlation CASSCF/CASPT2 calculations

Transition metal compounds The CASSCF method has been used extensively to study compounds containing transition metals. The choice of the active space is almost never trivial for such systems and must be closely related to the chemical process under study. The CASSCF method is usually used together with the CASPT2 method (which will be described in detail below) to add dynamic correlation effects. That combination often makes it necessary to use a larger active spaces than one would need, for example, if one was combining CASSCF with MRCI calculations even if that is also non-trivial in many cases. 

The selection of the active space in CASSCF and CASPT2 calculations is by no means trivial. In the present case we assumed that (i) the selected active space should at least represent the valence bonding and antibonding tr orbitals in BMe, (ii) the corresponding CAS scheme must lead to correct dissociation products (B 2P, Me 2S) and should be well balanced along the potential curve up to the dissociation limit, (iii) the CAS scheme should account for as much dynamical correlation as possible, and (iv) the same active space can be employed for all molecules in the series, i.e., BCu, BAg, and BAu. 

---