Geometric dependence of exchange bias in tilted three-dimensional CoFe/IrMn microwires

TL;DR

This study demonstrates exchange-biased 3D CoFe/IrMn microwires fabricated by combining two-photon lithography with magnetron sputtering. Using dark-field MOKE, the authors show that both the exchange-bias field $H_{EB}$ and the coercive field $H_c$ increase with wire inclination due to a geometry-induced reduction in the effective FM thickness $t_G^{FM}$, yielding a linear dependence of $H_{EB}$ on $1/t_G^{FM}$. The AF thickness plays a secondary role, with a critical AF thickness around $t_{cr}^{AF} oughly 3~ ext{nm}$ beyond which EB saturates; surface roughness and shadowing on inclined, rough surfaces can suppress $H_{EB}$ by generating thickness inhomogeneities. The results indicate that high-quality 3D interfaces can be achieved with smooth scaffolds, and that geometric thickness variations are a dominant factor in tailoring EB in 3D spintronic architectures, opening avenues for 3D magnetic devices and further studies of interfacial phenomena.

Abstract

The exchange bias (EB) effect, arising from interfacial coupling between ferromagnetic (FM) and antiferromagnetic (AF) layers, induces a unidirectional magnetic anisotropy and underpins a wide range of spintronic functionalities. Extending the EB effect to three-dimensional (3D) architectures enables investigation of interfacial coupling in non-planar structures, which is a key step toward realizing spintronic functionalities beyond planar systems. Achieving this requires the fabrication of FM/AF bilayers with smooth interfaces and well-defined thicknesses on non-planar scaffolds, together with suitable characterization methods. In this work, we realize exchange-biased 3D FM/AF microwires by combining two-photon lithography with magnetron sputtering. CoFe/IrMn bilayers are deposited on microwire scaffolds with inclination angles of 0 deg, 30 deg, 45 deg relative to the substrate, and their magnetization reversal is probed using dark-field magneto-optical Kerr effect (DF-MOKE) magnetometry. We find that the EB and coercive fields vary in a characteristic way with the inclination angle, consistent with the systematic reduction in film thickness expected from inclined directional deposition. In addition, the EB magnitude is influenced by the combined effects of surface roughness of non-planar geometries and the directional growth of the bilayer, highlighting the importance of 3D scaffold surface quality for integrating magnetic multilayers. These results provide insight into the growth and magnetic behavior of sputter-deposited magnetic multilayers with functional interfaces on 3D geometries.

Figures (4)

• Figure 1: Overview of the exchange-biased 3D CoFe/IrMn microwires.(a) Schematic of the 3D wires with inclination angles of 0°, 30° and 45° coated with a FM/AF bilayer. Note the two coordinate systems: global ($x$-$y$-$z$) and local ($x'$-$y'$-$z'$). (b) Scanning electron microscopy (SEM) images of the 3D CoFe/IrMn microwires (length 20 $\mu$m and width 1 $\mu$m) of inclination angles of 0°, 30° and 45°. The inset presents the deposited layer stack on the 3D wire scaffolds. (c) Schematic illustration of the sputter-deposition geometry with respect to the wire inclined at an angle $\theta$, highlighting the geometric thickness of the layer, $t_{G}$, compared to the nominal thickness on the substrate $t_{0}$. (d) Expected reduction in the layers (FM and AF) thickness as a function of inclination angle, $\theta$, for a highly directional deposition process. (e) Illustration of expected increase in the $H_{EB}$ with inclination angle $\theta$, arising due to the reduced geometric FM thickness $t_{G}^{FM}$, based on purely geometrical deposition arguments. The top $x$-axis corresponds to the inverse normalized $t_{G}^{\mathrm{FM}}$.
• Figure 2: Fabrication of exchange-biased 3D CoFe/IrMn microwires.(a) Schematic illustration of the TPL fabrication of 3D microwire scaffolds on a glass substrate, polymerization occurring at the focus point of the laser within the photoresist. (b) Schematic of the magnetron sputter deposition performed in the presence of an in-plane magnetic field applied along $x$-axis (see Fig. \ref{['fig:main_figure']}a for the coordinate systems). (c) Illustration of the resulting possible CoFe magnetization orientation in planar film and the 3D wire as a result of deposition in the presence of a magnetic field, followed by the IrMn deposition. The inset shows schematically the possible spin orientations of the CoFe/IrMn bilayer. (d-f) SEM images of the 3D microwires with inclinations of 0°, 30° and 45° (with respect to the substrate) after the deposition of the stack.
• Figure 3: Probing the simultaneous magnetization reversal of CoFe5.5/IrMn14 planar film and 3D microwires using DF-MOKE.(a) Schematic of the DF-MOKE setup. (b) Magnetization reversal loops of planar film and wires of length 20 $\mu$m and width 1 $\mu$m. The magnetic field $H_{x'}$, is applied along $x'$-direction (see Fig. \ref{['fig:main_figure']}a for coordinate systems). (c) Exchange bias and (d) coercivity plotted as a function of the wire's inclination angle, $\theta$. Solid symbols represent the data of the planar film, while open symbols represent the data of the individual wires. The top $x$-axes of both panels correspond to $1/t_{G}^{\mathrm{FM}}$. In panels (c) and (d), the same symbols are used and dashed lines are used to connect the data points.
• Figure 4: Exchange bias and coercivity as a function of wire geometry and surface roughness.(a) EB and (b) Coercivity as a function of inverse geometric FM thickness ($1/t_{G}^{\mathrm{FM}}$) for FM- and AF-series. Solid symbols correspond to planar film data, while open symbols represent wire data. Dashed lines indicate linear fits to the wire data at the three inclination angles, whereas the solid black line shows the linear fit to the planar film data in the FM series, with the dotted black line representing its extrapolation to lower $t_{0}^{\mathrm{FM}}$ values. $\sigma_{\mathrm{Planar\ films/Wires}}^{\mathrm{CoFe}\mathit{X}/\mathrm{IrMn}\mathit{Y}}$ (in $\mathrm{mT\,nm}$) denotes the slope obtained from the linear fit to the "Planar films/Wires" data points of the $\mathrm{CoFe}\mathit{X}/\mathrm{IrMn}\mathit{Y}$ bilayer. The labels $0^\circ$, $30^\circ$ and $45^\circ$ associated with the wire data points indicate the corresponding wire inclination angles. (c) SEM images of 0° wires (first column) for (1) CoFe5.5/IrMn14, (2) CoFe7/IrMn14 and (3) CoFe5.5/IrMn5, with corresponding AFM topographies (second column) acquired at indicated marker positions. The height color scale is shared across AFM images. The third column shows magnified views of 45° wires. (d) Power spectral density (PSD) vs. spatial frequency $k$, with planar films shown as solid lines and wires as dashed lines. (e) Slopes, extracted from linear fits of $H_{\mathrm{EB}}$ versus $1/t_{G}^{\mathrm{FM}}$, are plotted as a function of RMS roughness. Roughness data correspond to the planar geometries: the $0^\circ$ wire and the average of the $\mathrm{CoFe}\mathit{X}/\mathrm{IrMn}14$ planar films, with the standard deviation from this averaging used as the error bar. Marker symbols and colors, as indicated in the legend at the top, are consistent across all panels.