L10 FePt, multilayer thin films, solid state reaction, TEM, magnetic recording


FePt films with the L10 phase have potential applications for magnetic recording and permanent magnets due to its high magnetocrystalline anisotropy energy density. Heat treatment of n multilayer films is one approach to form the L10 FePt phase through a solid state reaction. This thesis has studied the diffusion and reaction of n multilayer films to form the L10 FePt phase and has used this understanding to construct exchange spring magnets. The process-structure-property relations of n multilayer films were systematically examined. The transmission electron microscopy (TEM) study of the annealed multilayers indicates that the Pt layer grows at the expense of Fe during annealing, forming a disordered fcc FePt phase by the interdiffusion of Fe into Pt. This thickening of the fcc Pt layer can be attributed to the higher solubilities of Fe into fcc Pt, as compared to the converse. For the range of film thickness studied, a continuous L10 FePt product layer that then thickens with further annealing is not found. Instead, the initial L10 FePt grains are distributed mainly on the grain boundaries within the fcc FePt layer and at the Fe/Pt interfaces and further transformation of the sample to the ordered L10 FePt phase proceeds coupled with the growth of the initial L10 FePt grains. A comprehensive study of annealed n films is provided concerning the phase fraction, grain size, nucleation/grain density, interdiffusivity, long-range order parameter, and texture, as well as magnetic properties. A method based on hollow cone dark field TEM is introduced to measure the volume fraction, grain size, and density of ordered L10 FePt phase grains in the annealed films, and low-angle X-ray diffraction is used to measure the effective Fe-Pt interdiffusivity. The process-structure-properties relations of two groups of samples with varying substrate temperature and periodicity are reported. The results demonstrate that the processing parameters (substrate temperature, periodicity) have a strong influence on the structure (effective interdiffusivity, L10 phase volume fraction, grain size, and density) and magnetic properties. The correlation of these parameters suggests that the annealed n multilayer films have limited nuclei, and the subsequent growth of L10 phase is very important to the extent of ordered phase formed. A correlation between the grain size of fcc FePt phase, grain size of the L10 FePt phase, the L10 FePt phase fraction, and magnetic properties strongly suggests that the phase transformation of fccL10 is highly dependent on the grain size of the parent fcc FePt phase. A selective phase growth model is proposed to explain the phenomena observed. An investigation of the influence of total film thickness on the phase formation of the L10 FePt phase in n multilayer films and a comparison of this to that of FePt co-deposited alloy films is also conducted. A general trend of greater L10 phase formation in thicker films was observed in both types of films. It was further found that the thickness dependence of the structure and of the magnetic properties in n multilayer films is much stronger than that in FePt alloy films. This is related to the greater chemical energy contained in n films than FePt alloy films, which is helpful for the L10 FePt phase growth. However, the initial nucleation temperature of n multilayers and co-deposited alloy films was found to be similar. An investigation of L10 FePt-based exchange spring magnets is presented based on our understanding of the L10 formation in n multilayer films. It is known that exchange coupling is an interfacial magnetic interaction and it was experimentally shown that this interaction is limited to within several nanometers of the interface. A higher degree of order of the hard phase is shown to increase the length scale slightly. Two approaches can be used to construct the magnets. For samples with composition close to stoichiometric L10 FePt, the achievement of higher energy product is limited by the average saturation magnetization, and therefore, a lower annealing temperature is beneficial to increase the energy product, allowing a larger fraction of disordered phase. For samples with higher Fe concentration, the (BH)max is limited by the low coercivity of annealed sample, and a higher annealing temperature is beneficial to increase the energy product.


If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at

Graduation Date



Coffey, Kevin


Doctor of Philosophy (Ph.D.)


College of Engineering and Computer Science


Mechanical, Materials and Aerospace Engineering

Degree Program

Materials Science and Engineering








Release Date

December 2008

Length of Campus-only Access


Access Status

Doctoral Dissertation (Open Access)