New materials for spin electronics

Abstract

Magnetic components are currently widely used in electronic devices, especially in the field of data storage. The need for downscaling the lateral dimensions of every component following Moore's law brings new challenges and will preclude the use of the currently employed materials and structures. The study presented is concerned with the investigation of new materials for future use in spin electronics devices. The first part focuses on full and half Heuser alloys Mn3Ga, Mn2Ga and the unprecedented hybrid Mn2RuxGa. We investigated Mn3Ga in the D022 cubic structure with large tetragonal distortion using x-ray absorption and magnetic circular dichroism (XAS/XMCD). We found that two different magnetic moments are present: the main component shows high perpendicular magnetic anisotropy (PMA), while a second, smaller one is oriented in the plane and posses no anisotropy. This configuration is due to the canted magnetic moment on one sublattice. The decomposition of the total magnetisation into the site-specific moments has been possible through the measurement of samples of Mn3Ga, Mn2.5Ga and Mn2Ga at several angles of incidence of the x-rays. We then investigated the properties of the first half-metallic compensated ferrimagnet, Mn2RuxGa. The two Mn sublattices are antiferromagnetically coupled and their moments cancel out almost perfectly. At the same time they are crystallographically inequivalent and only electrons with a definite spin participate in the electric conduction, making Mn2RuxGa a half-metal. The transport and magnetic properties have been studied by anomalous Hall effect (AHE) and by XAS/XMCD. We found that it is possible to tune the magnetic moment and the spin polarisation by Ru concentration x and tetragonal distortion c=a. This material is free from stray fields and unaffected by external magnetic perturbations. In addition, the very high spin polarisation should also lead to high magnetoresistance ratio if used in a magnetic tunnel junction device. Finally, we studied the spin-orbit torques of heavy metals with high spin-orbit coupling in order to investigate the possibility of the current-induced switching of an adjacent magnetic element. These torques arise from the spin Hall effect and from the Rashba effect. In particular we analysed materials such as W and Pt employing different methods in order to determine the validity of each method. The figure of merit used is the spin Hall angle, which measures the efficiency of the spin-orbit interaction in converting charge current into spin current. The literature values for the spin Hall angle of these metals vary of orders of magnitude, depending on the technique used. We found that the spin Hall angle of Pt assumes approximatively the same value regardless of the technique used, while the determination of the spin Hall angle for Ta and W gives very different results depending on the technique. Some types of measurement, such as the spin pumping from a ferromagnet and detection by inverse spin Hall effect or the harmonic Hall voltage measurement, are not reliable with these materials. We found an unusually high planar Hall resistance in W that could be at the origin of these unreliabilities.