Mr. Nykwest,

CFT represents an idealized description of the electronic structure of the coordination compounds where the ligands are described as point charges. Actually, looking at the data on very simple inorganics in solution like [Cu(H2O)4]2+ the theory explains well the electronic properties. However in context of complex electronic effects, complex ligands, specific experimental conditions and more, CFT fails to produce reliable explanation of the phenomena.

In this context you could provide your view of the phenomena according your results.

Hi Erik,

I think it can be useful if you compare the states between Al and Fe. The average oxidation state of Al in the Al2O3 is 3+, that means it shares its three valence electrons to form the bonds with the O atoms. Fe has a [Ar]3d6 4s2, if its in the same environment it is possible Fe looses 3 electrons too, but it will not gain a closed shell configuration as Al. That is why you have occupied intra-band states, and more empty d states.

I hope it helps you.

Thank you both for your responses. The purpose of using CFT was to predict the number of unpaired electrons in the doped system before performing DFT computations. My major concern with these results was that, to first order, CFT should be qualitatively similar and having the eg levels lower in energy than the t2g levels seemed to violate physics. However, after searching research gate and much reflection I believe the problem is VASP is decomposing the states using a different coordinate system than I expected, which would explain why many of the state have mixed decomposition character. Unfortunately I have how to change the coordinate system without creating a new reoriented supercell and redoing all of the calculations.

I may have to post a different question later on.