Band structure calculations of CrO2 carried out in the LSDA+U approach reveal a clear picture of the physics behind the metallic ferromagnetic properties. Arguments are presented that the metallic ferromagnetic oxide CrO2 belongs to a class of materials in which magnetic ordering exists due to double exchange (in this respect CrO2 turns out to be similar to the CMR manganates). It is concluded that CrO2 has a small or even negative charge transfer gap which can result in self-doping. Certain experiments to check the proposed picture are suggested. 75.50.Ss, 75.30.Et There has been a revival in interest in 3d transition metal oxides during the last decade. This was initially stimulated by the discovery of High-T c superconductivity in complex copper oxides, and more recently by the active study of the colossal magnetoresistance manganates (CMR) La 1−x (Ca,Sr) x MnO 3 . These latter systems in the most interesting composition range are metallic ferromagnets. This is interesting in itself because ferromagnetic ordering is rare among the oxides: most of them are antiferromagnetic or ferrimagnetic with dominating antiferromagnetic interactions. We need to meet certain special conditions to stabilize ferromagnetism. One of the main mechanisms invoked to explain ferromagnetic ordering in these systems is the double exchange mechanism [1], although it is not the only one [2].Another very well known ferromagnetic and metallic compound is chromium dioxide CrO 2 widely used in magnetic recording tapes. In its formal 4+ valence state Cr has two 3d electrons in t 2g orbitals which in a simple picture of strong correlations would suggest a Mott insulating-like ground state with S=1 local moments and most likely antiferromagnetic spin order. This seems to be about as far from the actual observed properties as one can get. In this letter we will address this problem using band structure methods supplemented with local Coulomb and exchange interactions (LSDA+U, [3]) interpreted in terms of local electronic configurations. We will argue that the d electrons can be divided into a localized "core" of spin 1 2 and an itinerant d electron propagating through these "cores" resulting in a double exchangelike mechanism for the ferromagnetic order much as in the manganates. We also show that strong electron correlation effects do not in this case lead to an insulating ground state for reasonable values of the d-d Coulomb interaction and perfect stoichiometry because CrO 2 should be viewed as a small or even negative charge transfer gap [4] material in the Zaanen-Sawatzky-Allen (ZSA) scheme[5] quite unlike the parent CMR material LaMnO 3 which is a Mott-Hubbard insulator. This leads for CrO 2 quite naturally to a phenomenon which could be referred to as self-doping resulting in a non-integral 3d band occupa-CrO 2 is a ferromagnetic metal with a saturation magnetic moment of 2.00 µ B and a low temperature resistivity with a nearly T 2 temperature dependence [7]. Band structure calculations in LSDA [8] explain this behaviour as that of a...

An explanation is given for the charge order, orbital order and insulating state observed in halfdoped manganese oxides, such as Nd 1/2 Sr 1/2 MnO3. The competition between the kinetic energy of the electrons and the magnetic exchange energy drives the formation of effectively one-dimensional ferromagnetic zig-zag chains. Due to a topological phase factor in the hopping, the chains are intrinsically insulating and orbital-ordered. Most surprisingly, the strong Coulomb interaction between electrons on the same Mn-ion leads to the experimentally observed charge ordering. For doping less than 1/2 the system is unstable towards phase separation into a ferromagnetic metallic and charge-ordered insulating phase. *

No abstract

No abstract

We consider the double-exchange for systems in which doped electrons occupy degenerate orbitals, treating the realistic situation with double degenerate eg orbitals. We show that the orbital degeneracy leads in general to formation of anisotropic magnetic structures and that in particular, depending on the doping concentration, the layered magnetic structures of the A-type and chain-like structures of the Ctype are stabilized. The phase-diagram that we obtain provides an explanation for the experimentally observed magnetic structures of some over-doped (electron-doped) manganites of the type Nd1−xSrxMnO3, Pr1−xSrxMnO3 and Sm1−xCaxMnO3 with x > 0.5.The double-exchange (DE) model [1][2][3] is one of the main models of ferromagnetism in metallic systems. The interest in this model was renewed recently in connection with the active study of the manganites with colossal magneto-resistance (CMR) [4]. Although it is not yet clear whether DE alone can explain the CMR in manganites [5], and there are some problems with the model itself (e.g. the tendency to phase separation [7][8][9][10]), the double-exchange mechanism is a necessary ingredient for the theoretical understanding of the CMR-effect [5,6] and remains the main explanation at least of the presence of the ferromagnetic state in doped manganites of the typeMost experimental data in the study of the CMRmaterials are obtained for the hole-doped manganites (x < 0.5) as it is usually in this doping range that CMR is observed. Recently, however, interesting data appeared on the properties of over-doped (or electron-doped) manganites (x > 0.5). Theoretically, in a conventional DE model one expects qualitatively similar behavior for small x and for x ∼ 1. Experimentally, however, there exists a very strong asymmetry. The behavior of doped manganites for x < 0.5 and x > 0.5 is drastically different. In particular a very stable insulating stripe phase is observed for x > 0.5 with the period determined by doping [11], whereas for x < 0.5, depending on the exact composition, either the metallic ferromagnetic state is realized [4], or a charged-ordered state with a doubled unit cell [12], which, however, can be easily transformed into a ferromagnetic metal by a magnetic field [13]. Recently Maignan et al. [14,15] succeeded in obtaining the state similar in some respects to the one predicted by the DE model in the electron-doped region. In Ca 1−y Sm y MnO 3 they observe metallic-like behavior with unsaturated ferromagnetism for y = 7 − 12% with indications that the state is not really ferromagnetic but has a more complicated, possibly canted, magnetic order.Other recent data [16][17][18][19] show that the system Nd 1−x Sr x MnO 3 is ferromagnetic for 0.25 < x < 0.5, A-type antiferromagnetic ordered for 0.5 < x < 0.6 and that C-type antiferromagnetic order is realized for 0.6 < x < 0.8. In Pr 1−x Sr x MnO 3 a A-type magnetic structure exists for 0.5 < x < 0.7. All these results show again that the behavior of the hole-doped and electrondoped manganites is indeed very differen...