Stacking-Order Effect on Spin-Orbit Torque, Spin Hall Magnetoresistance, and Magnetic Anisotropy in ${\mathrm{Ni}}_{81}{\mathrm{Fe}}_{19}–{\mathrm{Ir}\mathrm{O}}_2$ Bilayers

The 5d transition-metal oxides are an intriguing platform to demonstrate efficient charge-to-spin-current conversion due to a unique electronic structure dominated by strong spin-orbit coupling. Here, we report on the stacking-order effect of spin-orbit torque (SOT), spin-Hall magnetoresistance, and magnetic anisotropy in bilayer ${\mathrm{Ni}}_{81}{\mathrm{Fe}}_{19}$-5d iridium oxide, ${\mathrm{Ir}\mathrm{O}}_2$. While all ${\mathrm{Ir}\mathrm{O}}_2$ and $\mathrm{Pt}$ control samples exhibit large dampinglike SOT generation, stemming from the efficient charge-to-spin-current conversion, the magnitude of the SOT is larger in the ${\mathrm{Ir}\mathrm{O}}_2$($\mathrm{Pt})$ bottom sample than in the ${\mathrm{Ir}\mathrm{O}}_2$($\mathrm{Pt})$ top one. The fieldlike SOT has an even more significant stacking-order effect, resulting in an opposite sign in the ${\mathrm{Ir}\mathrm{O}}_2$ samples in contrast to the same sign in the $\mathrm{Pt}$ samples. Furthermore, we observe that the magnetic anisotropy energy density and the anomalous Hall effect are increased in the ${\mathrm{Ir}\mathrm{O}}_2$($\mathrm{Pt})$ bottom sample, suggesting enhanced interfacial perpendicular magnetic anisotropy. Our findings highlight the significant influence of the stacking order on spin transport and the magnetotransport properties of $\mathrm{Ir}$ oxide-ferromagnet systems, providing useful information for the design of SOT devices, including 5d transition-metal oxides.

Figure 1

(a) Top, cross section of bilayer films labeled as ${\mathrm{Ir}\mathrm{O}}_2$ top (T), Py(3)/${\mathrm{Ir}\mathrm{O}}_2$(10) and ${\mathrm{Ir}\mathrm{O}}_2$ bottom (B), ${\mathrm{Ir}\mathrm{O}}_2($10)$/$Py(3). Bottom, patterned device structure with two Hall bars, electrical detection, and coordinate system. (b) Saturation magnetization and (c) sheet resistance for samples with ${\mathrm{Ir}\mathrm{O}}_2\text{−}T$ and ${\mathrm{Ir}\mathrm{O}}_2\text{−}B$, including $\mathrm{Pt}$ control samples of Py(3)/$\mathrm{Pt}$(6) and $\mathrm{Pt}($6)/Py(3).

Figure 2

(a) Hall resistance of ${\mathrm{Ir}\mathrm{O}}_2\text{−}T$ sample as a function of perpendicular magnetic field at $\theta =0^{\circ }$. (b) Perpendicular magnetic anisotropy energy density and (c) anomalous Hall resistance for all samples extracted from measurements, similar to that shown in (a).

Figure 3

(a) Illustration of rotation planes. (b) First-harmonic longitudinal resistance of ${\mathrm{Ir}\mathrm{O}}_2\text{−}T$ sample measured by rotating samples with a fixed external field of 1.3 T. Black line is the result of fitting using Eq. (1). (c)–(e) Magnetoresistance expressed as percentage in three planes for all samples.

Figure 4

(a) First-harmonic Hall resistance $(R_H^{1\omega })$ of ${\mathrm{Ir}\mathrm{O}}_2\text{−}T$ sample measured at 0.1 T (red open circle) and 0.5 T (blue open square). Black solid curve is fitting of data using Eq. (2). (b) Second-harmonic Hall resistance $(R_H^{2\omega })$ of ${\mathrm{Ir}\mathrm{O}}_2\text{−}T$ sample measured at 0.1 and 0.5 T for raw data (top panel). Solid curves are fits to data using Eq. (4). Separated $\cos \varphi$ and $2\mathrm{co}{\mathrm{s}}^3\varphi \text{ - }\cos \varphi \mathrm{components}$ from $R_H^{2\omega }$ indicate the DL contribution and FL contribution, respectively. (c) Obtained $R_{\text{DL + }\mathrm{\nabla }T}$ as a function of $1/(B_{\mathrm{ext}}+B_k^{\mathrm{eff}})$ and (d) $R_{\text{FL + Oe}}$ as a function of 1/B_ext for the ${\mathrm{Ir}\mathrm{O}}_2\text{−}T$ sample. Red solid lines represent linear fits. (e) B_DL/J and (f) B_FL/J for all samples. J is the current density flowing in the nonmagnet layer normalized to 10¹¹ A/m².

Figure 5

(a) Dampinglike spin-orbit-torque efficiency evaluated by harmonic Hall and spin-Hall magnetoresistance measurements for all samples. (b) Fieldlike spin-orbit-torque efficiency for all samples.