Spintronics is a fast-developing field which makes use of the two spin states of the electron, and has the potential for more efficient, robust, and faster microelectronic devices. Thin films of rare-earth iron garnets, a class of insulating ferrimagnetic oxides, are particularly well suited to this application as the anisotropy, magnetization, magnetostriction, and damping can be easily controlled through selection of rare-earth ion and substrate. Previous work on garnets has focused on epitaxial single-crystal films grown on garnet substrates, which are expensive and not of commercial importance. Thus, it is of interest to grow nanometer scale thin films of garnets as polycrystalline layers on non-garnet substrates with perpendicular magnetic anisotropy.
In this thesis, the growth of polycrystalline thin films of rare-earth iron garnets with controllable anisotropy and spin transport properties comparable to single crystal films is reported. Perpendicular magnetic anisotropy, which is essential for efficient manipulation of the magnetization through spin-orbit torque injection from an adjacent conductive layer, is achieved via control of the magnetoelastic anisotropy from thermal expansion mismatch between the film and substrate for europium iron garnet/quartz and dysprosium iron garnet/silicon. Heterostructures with a platinum overlayer allow investigation of the spin Hall magnetoresistance, which indicates a high degree of spin transparency at the interface. Next, a novel heterostructure is developed that allows for the growth of ultra-thin (< 10 nm) polycrystalline garnet films through the use of a nonmagnetic seed layer and diffusion barrier. The magnetic proximity effect in heavy metals is also investigated across the magnetic compensation temperature of dysprosium iron garnet, demonstrating exchange coupling behavior different from that of metallic magnets and validating the importance of a sharp, contamination free interface in these materials. Finally, a molecular field coefficients model is modified to account for non-ideal stoichiometry and site occupancy in the garnet crystal structure. The model is used to explain discrepancies between the bulk and thin film magnetic compensation temperatures. The work demonstrated here outlines the potential for integration of magnetic insulators into next-generation spintronic devices.
Ph.D.