Dear Adf'ers,

                I have a basic question regarding the calculation of dissociation curves
using Adf.
                Let's have, for example, the carbon dimer in its singlet ground state.
There's no problem here because using a broken symmetry approach (C_{\infty
v} instead of D_{\infty h}) together with the fragoccupations (the
fragments are in fact atoms in C_{\infty v} symmetry, too) and
modifystartpotential keywords things go well. I have two separated carbon
atoms in a triplet state, one with two spin-up electrons and the other with
two spin-down ones. Moreover, D_e and R_e are in very good agreement with
other post-HF calculations.
                However, if I'm interested in the very_close (\Pi) triplet state of C_2 I
don't know how to carry out the calculations. My problem arises because
that system dissociates into two carbon atoms with four total unpaired
electrons (or zero total unpaired electrons if you prefer). How can I
"link" the triplet state of C_2 in the bonding region to the quintet (or
singlet) state in the dissociation zone?
                I understand that this problem can be worked out at a post-HartreeFock
level using a MR (CAS) wavefunction to deal with the static correlation
component. It is not clear at all to me how to face that at the
monodeterminant level of DFT.
                Thanks for your interest.
                Regards to all,

                                                Reinaldo
Received on 2000-07-12 14:23:20