Exchange paths were investigated for unidirectional exchange coupled 40 mn Nis,Fe,dSO mn NiO films by performing several field cooling experiments. Our experimental data were very consistent with the assumed existence of a variety of exchange paths. Each exchange path seemed to produce its own local unidirectional anisotropy and different local blocking temperature. The measureable exchange coupling could be described as consisting of the sum of the respective exchange paths, each with its own local blocking temperature. On the other hand, an observed blocking temperature of about 230 "C was determined from the exchange paths having the highest local blocking temperature. The local blocking temperatures were thought to be widely distributed, ranging from room temperature to about 230 "C, and the maximum existence probability was most likely at about 215 "C. This indicated that the exchange paths having the local blocking temperature of 215 "C made the largest contribution to the exchange coupling field at room temperature. According to cross sectional transmission electron microscopy observations, this variety of exchange paths was caused by inhomogeneous N&Fe,,-NiO interfaces associated with inter-facial disorder and fluctuating atomic arrangement.

Single-crystal Fe16N2 films have been grown epitaxially on Fe(001)/InGaAs(001) and InGaAs(001) substrates by molecular beam epitaxy (MBE). Saturation flux density Bs of Fe16N2 films has been demonstrated to be 2.8–3.0 T at room temperature, which is very close to the value obtained by Kim and Takahashi using polycrystalline evaporated Fe–N films. Temperature dependence of Bs has been measured. Bs changed with temperature reversibly up to 400 °C, while beyond 400 °C, Bs decreased irreversibly. X-ray diffraction showed that Fe16N2 crystal is stable up to 400 °C, while beyond 400 °C, Fe16N2 dissolves into Fe and Fe4N, and also some chemical reactions between Fe16N2 and the substrate occurs. This caused the temperature dependence of Bs mentioned above. From the temperature dependence of Bs up to 400 °C, the Curie temperature of Fe16N2 is estimated to be around 540 °C by using the Langevin function. The above mentioned Bs of 2.9 T at room temperature and 3.2 T at −268 °C corresponded to an average magnetic moment of 3.2μB per Fe atom and 3.5μB, respectively. These values of the magnetic moment of Fe atoms are literally giant, far beyond the Slater–Pauling curves. The origin of the giant magnetic moment has been discussed based on the calculation carried out by Sakuma. However, there was a significant disagreement between experimental values and calculated ones, so the origin remained to be clarified. Also, magneto-crystalline anisotropy of Fe16N2 films has been investigated.

Single-phase, single-crystal Fe16N2(001) films and Fe-11 at. %N martensite films of 200–900 Å thickness have been epitaxially grown on In0.2Ga0.8As(001) substrates by evaporating Fe in an atmosphere of mixed gas of N2 and NH3, followed by annealing. The saturation magnetizations 4πMs’s for Fe16N2 and Fe-N martensite films have been measured to be around 29 and 24 kG at room temperature, respectively, and almost constant in the above thickness range by using a vibrating sample magnetometer. 4πMs for Fe-N martensite films has been increased with ordering of N atoms caused by annealing and finally reached around 29 kG for Fe16N2. Mössbauer spectra have been measured for those films. The spectrum for Fe-N martensite films was a superposed one with hyperfine fields of 360, 310, and 250 kOe, similar to those previously reported for martensite. While the spectrum became simpler with ordering, finally reaching a single hyperfine field of 330 kOe for Fe16N2. 4πMs of 29 kG for Fe16N2 (3.2 μB/Fe atom) and 4πMs of 24 kG for martensite (2.6 μB/Fe atom) has not been explained based on the conventional band theory of 3d metal magnetism. Behaviors of Mössbauer spectra could not be understood based on the conventional concept either. Thus a new physical concept is likely to be needed for clarification of giant magnetic moments and Mössbauer spectra for Fe16N2 and Fe-N martensites.

No abstract