I hate to be the guy that always disagrees with people (
) but I am sure that when cations are formed, they lose their outer most electron i.e. the electron with the highest value for n (i.e the number of the shell) and
then the highest l value (s, p, d, f).
So I would say rleung is right to believe the electron configuration of the Co
2+ ion is [Ar]3d
7 because the two 4s electrons are lost first. I can't imagine any element losing their 3-something electrons before their 4-something electrons- same goes for 4&5 etc.
In fact in the case of Co
2+, I think [Ar]3d
7 would be stabler than the other one because when it is charged you have to think about the ligands bonding to the cobalt, and the large amount of repulsion between these and the theoretical 4s electrons.
3d
5 wouldn't be any more stable in the ion because I think it would be unlikely that one electron would occupy each orbital anyway, since the difference in energy levels between the orbitals; I imagine there would be 4 or 5 electrons occupying the lowest 3 energy orbitals rather than 3, assuming it is d
2sp
3 hybridised, so 3d
7 wouldn't be any worse.