I am having problems understanding the construction of MO diagrams in octahedral transition metal complexes within ligand field theory (LFT) when the metal center is asumed as an cation:
I am completely fine understanding the MO diagram when the metal is neutral and thus the (say) 3d, 4s, 4p orbital energies are above the ligand orbitals as shown e.g. here:
http://chemed.chem.purdue.edu/genchem/topicreview/bp/ch12/ligand.php
However, when the metal becomes cationic, wouldn't the metal orbitals be significantly lower in energy?
In textbooks, you find quite a few pictures like these: 
However, when I perform some quantum chemical calculations (it does not really matter if based on Density Functional Theory or Hartree-Fock), the metal orbitals drop significantly below the ligand orbitals (as you would expect).
With such an energy difference, the metal and ligand orbital should not be able to interact properly as they don't match energetically.
Consider for example orbital energies of [Fe(H$_2$O)$_6$]$^{2+}$ on BP86/SVP level of theory:
- HOMO [Fe(II,LS)(H$_2$O)$_6$]$^{2+}$ : -12.0997 eV
- HOMO [(H$_2$O)$_6$] : -4.7910 eV
- HOMO [Fe(II,LS)]$^{2+}$ : -25.9057 eV
Now, if Fe is assumed neutral, the problem more or less "vanishes" as the metal orbitals rise in energy and the diagram makes way more sense:
- HOMO [Fe(0)] : -5.1303 eV
However, this is not what the diagram shows.
Can you help in explaining the discrepancy?
Is the MO diagram making things way to easy?
Is the assumption of a cationic metal center non sensible?
Are orbital energies overrated and cannot be used for problems like these?
