“…In general, the slightly negative ∆g i arises from two causes: (i) the large contribution to the g factor due to the spin-orbit (SO) coupling parameter of the ligand via covalence effect in the case of a ligand having a large SO coupling parameter (where a two-SO-parameter model including both the SO coupling parameter of 3d n ion and that of ligand should be used [4,5]) and (ii) the large contribution to the g factor from the charge transfer (CT) mechanism in the case of the central 3d n ion having high valence state and hence low CT energy level [6] (thus, the influence of the CT excited state on the g factors of the ground state cannot be neglected and so both crystal field (CF) and CT mechanisms should be considered [7,8] [10]), the second cause may be important. In order to confirm the above opinion and to study the relative importance of the CT mechanism not only for the g factors, but also for zero-field splitting D and the change of the relative importance of these with the valence state of 3d 3 ions, in this paper we establish the complete high-order perturbation formulas including both CF and CT mechanisms for the EPR parameters of 3d 3 ions in tetragonal octahedral clusters. From these formulas, the EPR parameters g || , g ⊥ and D for Cr 3+ and Mn 4+ ions in anatase crystals are calculated.…”