Epitaxial FeGa/IrMn bilayers with exchange biases along the FeGa[100] and [110] directions are prepared on MgO(001) single crystal substrates by magnetron sputtering through controlling the orientation of the external field <i>in situ</i> applied during growth. The effect of the exchange bias orientation on the magnetic switching process and the magnetic switching field are studied. The X-ray <i>φ</i>-scan indicates that the FeGa layer is epitaxially grown with a 45° in-plane rotation on the MgO(001) substrate along the FeGa(001)[110] direction and the MgO(001)[100] direction. The measurements of the angular dependence of the ferromagnetic resonance field and the corresponding fitting to the Kittel equation show that the samples have a superposition of fourfold symmetric magnetocrystalline anisotropy <inline-formula><tex-math id="M4">\begin{document}$ {K}_{1} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M4.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M4.png"/></alternatives></inline-formula>, unidirectional magnetic exchange bias anisotropy <inline-formula><tex-math id="M5">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M5.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M5.png"/></alternatives></inline-formula>, and uniaxial magnetic anisotropy <inline-formula><tex-math id="M6">\begin{document}$ {K}_{\mathrm{u}} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M6.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M6.png"/></alternatives></inline-formula> with configuration of <inline-formula><tex-math id="M7">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}}//\left[100\right] $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M7.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M7.png"/></alternatives></inline-formula> or <inline-formula><tex-math id="M8">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}}//\left[110\right] $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M8.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M8.png"/></alternatives></inline-formula>. The combined longitudinal and transverse magneto-optical Kerr effect measurements show that sample with <inline-formula><tex-math id="M9">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}}//\left[100\right] $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M9.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M9.png"/></alternatives></inline-formula> exhibits square loops, asymmetrically shaped loops, and one-sided two-step loops in different external magnetic field directions. In contrast, the sample with <inline-formula><tex-math id="M10">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}}//\left[110\right] $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M10.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M10.png"/></alternatives></inline-formula> exhibits one-sided two-step and two-sided two-step loops as the magnetic field orientation changes. Because the <inline-formula><tex-math id="M11">\begin{document}$ {K}_{1} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M11.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M11.png"/></alternatives></inline-formula> is superimposed by <inline-formula><tex-math id="M12">\begin{document}$ {K}_{\mathrm{u}} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M12.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M12.png"/></alternatives></inline-formula> and <inline-formula><tex-math id="M13">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M13.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M13.png"/></alternatives></inline-formula>, the in-plane fourfold symmetry of the magnetic anisotropy energy is broken. The local minima are no longer strictly along the in-plane <inline-formula><tex-math id="M14">\begin{document}$ \left\langle{100}\right\rangle $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M14.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M14.png"/></alternatives></inline-formula> directions, but make a deviation angle which depends on the relative orientation and strength of magnetic anisotropy. A model based on the domain wall nucleation and propagation is proposed with considering the different orientations of <inline-formula><tex-math id="M15">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M15.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M15.png"/></alternatives></inline-formula>, which can nicely explain the change of the magnetic switching route with the magnetic field orientation and fit the angular dependence of the magnetic switching fields, indicating a significant change of domain wall nucleation energy as the orientation of <inline-formula><tex-math id="M16">\begin{document}$ {K}_{\mathrm{e}\mathrm{b}} $\end{document}</tex-math><alternatives><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M16.jpg"/><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="12-20220166_M16.png"/></alternatives></inline-formula> changes.
Read full abstract