Analysis of available experimental data shows that there exists a limited variety of shapes of temperature dependence of spontaneous magnetization. For most metallic ferromagnets the shape (as opposed to scale) of the M(s) versus T curve can be characterized by a single dimensionless parameter. A numerical description of the dependence M(s)(T) for a particular ferromagnetic material is thus reduced to evaluating three quantities: the saturation magnetization M(0) identical with M(s)(0), the Curie point T(C), and the shape parameter s. It is demonstrated that classical spin (S = infinity) dynamics fails to describe correctly either of the finite-temperature characteristics, T(C) or s.

It is demonstrated by means of density functional and ab-initio quantum chemical calculations, that transition metal -carbon systems have the potential to enhance the presently achievable area density of magnetic recording by three orders of magnitude. As a model system, Co 2 -benzene with a diameter of 0.5 nm is investigated. It shows a magnetic anisotropy in the order of 0.1 eV per molecule, large enough to store permanently one bit of information at temperatures considerably larger than 4 K. A similar performance can be expected, if cobalt dimers are deposited on graphene or on graphite. It is suggested that the subnanometer bits can be written by simultaneous application of a moderate magnetic and a strong electric field. PACS numbers: 31.15.es, 75.30.Gw, 75.75.+a Keywords: applications of density functional theory, magnetic anisotropy, magnetic properties of nanostructures 1 Long-term magnetic data storage requires that spontaneous magnetization reversals should occur significantly less often than once in ten years. This implies that the total magnetic anisotropy energy (MAE) of each magnetic particle should exceed 40 kT , 1 where k is the Boltzmann constant and T is the temperature. Among the elemental ferromagnets (Fe, Co, Ni, and Gd), cobalt metal shows the highest MAE, about 0.06 meV per atom in the hexagonal close packed structure. Thus, Co is the main ingredient of magnetic data storage materials at present. At room temperature, data loss due to fluctuations is avoided, if a Co grain contains not less than 40 k · 300 K/0.06 meV ≈ 15,000 atoms. In fact, the grain diameter of contemporary Co(Cr,Pt,SiO 2 ) recording media is close to 8 nm, each grain containing about 50,000 atoms and each bit being composed of some dozen grains.2 The grain size could be considerably reduced by using the intermetallic compounds FePt or CoPt with record MAE of almost 1 meV per atom in their structurally ordered L1 0 bulk phase. It is, however, hard to achieve the required ordered structure in nano-particles. 3Obviously, a further reduction of the bit size is primarily limited by the value of MAE per atom. Recent efforts to enhance this value were focused on single atoms or small clusters deposited on the surface of heavy metals. This approach combines two ideas: 4 Firstly, the magnitude of MAE is related to the size of the orbital moments. The latter are quenched for highly coordinated atoms but can be large if the coordination is low. Secondly, the magnitude of MAE is related to the strength of spin-orbit coupling which grows with atomic number. Considerable progress was achieved in this way by deposition of single Co atoms on a Pt surface, yielding a record MAE of 9 meV per Co atom. 5 Unfortunately, clusters of several Co atoms on Pt show a much smaller MAE per atom, roughly inversely proportional to the number of atoms. 5More recently, the magnetic properties of transition metal dimers came into the focus of interest. 6,7,8,9 Isolated magnetic dimers are the smallest chemical objects that possess a magnetic a...

No abstract

The temperature variation of the ͑100͒ and ͑010͒ neutron diffraction peak intensities, related only to the Nd magnetic moments, have been measured on a NdFeO 3 single crystal, at temperatures down to 70 mK. The ͑100͒ peak becomes noticeable below 25 K while the ͑010͒ peak only gives an appreciable contribution below 1 K. Above T N2 Ϸ1 K the ͑100͒ peak intensity is accounted for by the electronic magnetic moments polarized by the Nd-Fe exchange field. Near T N2 a change of slope is observed in the temperature dependence of the ͑100͒ reflection intensity, demonstrating the crossover from the above polarization of Nd under the Nd-Fe exchange to proper long-range ordering due to Nd-Nd interaction. Below ϳ0.4 K another mechanism, polarization of Nd nuclear moments by hyperfine field, contributes to the intensity of the ͑100͒ and ͑010͒ peaks. A simple mean-field model explains consistently the observed temperature dependence of the diffraction intensities as well as earlier specific-heat data. The main feature of this model is allowance for Van Vleck susceptibility, which appears to play an important role in the overall polarization of Nd. The values of the hyperfine field at the Nd nuclei H hf ϭ1.0Ϯ0.15 MOe and of the Nd electronic magnetic moment Nd ϭ0.9 B are deduced, the ratio H hf / Nd being the same as in other Nd compounds.

No abstract