| During the last decade, unexpected and unusual room temperature ferromagnetism has often been reported or claimed in systems that do not contain any of the necessary elementary bricks (magnetic ions). Among this broad family of often called «d0» compounds one finds, oxides such as HfO2, ZrO2, CaO, ZnO or CeO2, hexaborides such as CaB6 or even graphite, graphene and other C based systems. We propose here a theory (Hubbard model in the presence of correlated disorder) able to shed light on the underlying mechanisms and that provides the necessary conditions to control such non-magnetic defect induced ferromagnetism. Phase diagrams and detailed analysis with respect to the physical parameters (defect/vacancies concentration, carrier density, electron-electron interaction and disorder strength) are provided. In this real space approach, disorder and electronelectron correlations are treated simultaneously and on equal footings. Interestingly, it is found that high Curie temperature can effectively be reached for realistic values of the model parameters and for carrier density per non-magnetic defect within a certain window. The theory has also been able to predict that oxides such as Na, K doped HfO2, ZrO2 and TiO2 should be good candidate for high Tc ferromagnetism. This has indeed been confirmed by recent ab initio based calculations. |
|