Transmission Electron Microscopy for Characterization of Structures, Interfaces and Magnetic Moments in Magnetic Thin Films and Multilayers

Structural characterization is essential for the understanding of the magnetic properties of thin films and multilayers. In this thesis, both crystalline and amorphous thin films and multilayers were analyzed utilizing transmission electron microscopy (TEM). High resolution TEM and electron diffraction studies emphasize on the growth of amorphous Fe91Zr9 and Co68Fe24Zr8 on both Al2O3 and Al70Zr30 in multilayer structures by magnetron sputtering. The properties of the growth surfaces were found to strongly influence the formation of nano-crystallites of the magnetic material at interfaces. Field induced uniaxial magnetic anisotropy was found to be possible to imprint into both fully amorphous and partially crystallized Co68Fe24Zr8 layers, yielding similar magnetic characteristics regardless of the structure. These findings are important for the understanding of both growth and magnetic properties of these amorphous thin films. As magnetic systems become smaller, new analysis techniques need to be developed. One such important step was the realization of electron energy-loss magnetic circular dichroism (EMCD) in the TEM, where information about the ratio of the orbital to spin magnetic moment (mL/mS) of a sample can be obtained. EMCD makes use of angular dependent inelastic scattering, which is characterized using electron energy-loss spectroscopy. The work of this thesis contributes to the development of EMCD by performing quantitative measurements of the mL/mS ratio. Especially, methods for obtaining energy filtered diffraction patterns in the TEM together with analysis tools of the data were developed. It was found that plural inelastic scattering events modify the determination of the mL/mS ratio, wherefore a procedure to compensate for it was derived. Additionally, utilizing special settings of the electron gun it was shown that EMCD measurements becomes feasible on the nanometer level through real space maps of the EMCD signal.