Imaging the Magnetic Spin Structure of Exchange-Coupled Thin Films

We have investigated the magnetic spin structure of a soft-magnetic film that is exchange-coupled to a hard-magnetic layer to form an exchange-spring layer system. The depth dependence of the magnetization direction was determined by nuclear resonant scattering of synchrotron radiation from ultrathin ${}^{\mathrm{57}}\mathrm{F}\mathrm{e}$ probe layers. In an external field a magnetic spiral structure forms that can be described within a one-dimensional micromagnetical model. The experimental method allows one to image vertical spin structures in stratified media with unprecedented accuracy.

Figure 1

Left: scheme of the spin structure in an exchange-spring bilayer consisting of a soft-magnetic layer (Fe) on a hard-magnetic layer (FePt) with uniaxial anisotropy. An external field $\overrightarrow{H}$ is applied perpendicular to the remanent magnetization $\overrightarrow{M}$ of the hard layer. Right: to image the resulting magnetic spiral, a tilted ultrathin probe layer of ${}^{\mathrm{57}}\mathrm{F}\mathrm{e}$ is deposited within the soft layer. The magnetic properties in depth $D$ of the film are probed by adjusting the transverse displacement $\Delta x$ of the sample relative to the scattering plane.

Figure 2

The $\mathrm{A}\mathrm{g}/\mathrm{F}\mathrm{e}/\mathrm{F}\mathrm{e}\mathrm{P}\mathrm{t}$ trilayer as an x-ray waveguide. (a) Depth dependence of the electric field intensity at an angle of incidence of ${\phi }_m=4.4\mathrm{m}\mathrm{r}\mathrm{a}\mathrm{d}$. At this angle, the radiation couples into a guided mode, forming a standing wave inside the Fe layer. This shows up as a minimum in the rocking curve of the nonresonant reflectivity (b) and as a pronounced peak in the flux of the resonantly scattered radiation (c), recorded within a time range from 20–160 ns after excitation.

Figure 3

(a) Modulus squared of the energy-dependent resonant strengths $F_{+1}(\omega )$ (solid line) and $F_{-1}(\omega )$ (dashed line) for the 14.4 keV transition in ferromagnetic ${}^{\mathrm{57}}\mathrm{F}\mathrm{e}$. (b) Hyperfine field distribution that was used to simulate the time spectra of the probe layer close to the $\mathrm{F}\mathrm{e}/\mathrm{A}\mathrm{g}$ interface.

Figure 4

Measured time spectra of grazing incidence reflection. Left: variation of the lateral positions $\Delta x$ in a perpendicular magnetic field of $B=160\mathrm{m}\mathrm{T}$. Right: variation of the external magnetic field at $\Delta x=10\mathrm{m}\mathrm{m}$, where the ${}^{\mathrm{57}}\mathrm{F}\mathrm{e}$ probe layer is in the center of the Fe layer. In each case the change in the beat pattern reflects the rotation of the magnetization direction. Solid lines are theoretical simulations according to Eq. (1). The rotation angles derived from the simulations are shown in Fig. 5.

Figure 5

Depth dependence of the spin rotation in the Fe layer for external magnetic fields of 160 mT ($•$), 240 mT ($▴$), and 500 mT ($\blacksquare$). Shaded regions mark the Ag capping layer and the FePt substrate, respectively. Solid lines are simulations according to the model outlined in the text. The dashed lines are simulations assuming a reduced exchange coupling in the first 3 nm of the Fe layer. The left inset shows the depth dependence of the Fe oxide phase, given in at. %. The right inset displays the rotation angle in the center of the Fe layer as a function of the external magnetic field with simulations for $A_s=1.0\times {10}^{-6}\mathrm{e}\mathrm{r}\mathrm{g}/\mathrm{c}\mathrm{m}$ (solid line) and $A_s=2.8\times {10}^{-6}\mathrm{e}\mathrm{r}\mathrm{g}/\mathrm{c}\mathrm{m}$ (dashed line).