Fig. 5(a) plots the Fe-2p XPS spectrum of the as-deposited and
annealed FePt films with Ta ¼ 650e800 C. The vertical dash lines
represent the metallic Fe and Fe2O3 binding energies as references.
Note that the samples were sputtered for 1 min to remove the
oxygen layers which were physically absorbed on the FePt surfaces
in the XPS chamber before measurement. Besides, the XPS
detectable depth (d) is determined by the formula of d ¼ 3Msina,
where theMand a are the inelastic mean-free paths (IMFP) and the
take off angles for photoelectrons (in this study, a ¼ 45). For the xray
beam with energy of 1 keV, the IMFP of the electrons of FePt is
about 1.3 nm [28]. Thus, the detectable depth is determined to be
around 2.8 nm. For the as-deposited FePt film, the peaks were
located on the metallic Fe core-level. After an annealing of 650 C,
the peaks remained unchanged, indicating that the FePt film wasnot surface-oxidized during RTA. As Ts was further increased to
700 C, the peaks started to shift toward Fe2O3 state, which confirms
that the formation of FeOx on the surface of the L10-FePt film.
For the FePt film RTA at 800 C, a further shift toward Fe2O3 state
was observed, owing to the deeper oxygen diffusion and formation
of thicker oxide layer. The XPS spectra of the FePt films annealed
700 C show convolution curves from Fe2O3 component and pureFePt component, implying an off-stoichiometric FeeO compound.
The convolution curves indicate that the thicknesses of the oxidation
layer should be much less than 2.8 nm.
To further understand the role of surface oxidation on (001)
preferred orientation, the FePt films was capped with an MgO film
(1 nm-thick) and then was rapid-annealed at 800 C. The MgO film
was used to prevent exposure to oxygen atmosphere during
annealing. As shown in Fig. 5(b), the sample exhibits an L10 structure
with S value of ~0.9. However, a poor perpendicular anisotropy
with a LOF value of 0.72 was obtained in the bilayer film. Furthermore,
we also examined the s of the bilayer film which exhibits a
much smaller value (1.5 GPa) than that of the samples with free
surface (3.3 GPa). Since the capping oxide layer was expected to
suppress the surface oxidation of the L10 FePt films during RTA, it
can be implied that surface oxidation of the L10 FePt films which
induces strong tensile stress probably play a very important role in
enhancing the perpendicular anisotropy. Further investigations are
required to fully understand surface-oxidation phenomena and
resultant surface stress of the FePt films during RTA