Room-Temperature Spin-Orbit Torque Switching Induced by a Topological Insulator

The strongly spin-momentum coupled electronic states in topological insulators (TI) have been extensively pursued to realize efficient magnetic switching. However, previous studies show a large discrepancy of the charge-spin conversion efficiency. Moreover, current-induced magnetic switching with TI can only be observed at cryogenic temperatures. We report spin-orbit torque switching in a TI-ferrimagnet heterostructure with perpendicular magnetic anisotropy at room temperature. The obtained effective spin Hall angle of TI is substantially larger than the previously studied heavy metals. Our results demonstrate robust charge-spin conversion in TI and provide a direct avenue towards applicable TI-based spintronic devices.

Figure 1

Structure and magnetic properties of the ${\mathrm{Bi}}_2{\mathrm{Se}}_3/\mathrm{CoTb}$ sample. (a) Schematic of SOT in ${\mathrm{Bi}}_2{\mathrm{Se}}_3/\mathrm{CoTb}$ heterostructure. The charge current (${\mathit{j}}_e$) generates spin accumulation ($-\mathit{\sigma }$) which is perpendicular to the current direction at the interface, and exerts a SOT on the magnetic moments ($\mathit{m}$). The SOT is proportional to $\mathit{m}\times (\mathit{\sigma }\times \mathit{m})$ and can be described by an effective field ${\mathit{H}}_{\mathrm{SO}}$, which is proportional to $\mathit{\sigma }\times \mathit{m}$. (b) Hysteresis loop of the out-of-plane magnetization measured by vibrating sample magnetometry (VSM). (c) Image of the Hall device with an illustration of the electrical measurement setup. The current is applied along the $x$ axis and the Hall voltage is detected in the $y$ direction. The width of the Hall bar is $\sim 4\mu \mathrm{m}$. (d) Anomalous Hall resistance vs out-of-plane magnetic field.

Figure 2

Room temperature SOT switching in ${\mathrm{Bi}}_2{\mathrm{Se}}_3/\mathrm{CoTb}$. Hall resistance is measured when sweeping a dc current under a bias field of 1000 Oe along (a) the positive and (b) negative $x$ axis. $j_e$ is the average current density inside the ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ layer.

Figure 3

Calibration of the SOT efficiency. (a) Hall resistance vs applied perpendicular field under positive and negative dc currents with the density of $1.0\times {10}^6\mathrm{A}/{\mathrm{cm}}^2$ through ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ and a bias field $H_x=+800\mathrm{Oe}$. The center shift corresponds to the SOT-induced effective field ($H_z^{\text{eff}}$). (b) $H_z^{\text{eff}}$ as a function of applied current under bias fields $H_x=0,\pm 200\mathrm{Oe}$. (c) SOT efficiency $\chi$ vs $H_x$. $\chi$ saturates at the field $H_x^{\text{sat}}$. The data shown in (a) to (c) are measured in the ${\mathrm{Bi}}_2{\mathrm{Se}}_3/\mathrm{CoTb}$ sample. (d) SOT efficiency $\chi$ vs $H_x$ in the ${(\mathrm{Bi},\mathrm{Sb})}_2{\mathrm{Te}}_3/\mathrm{CoTb}$ sample. Note that the magnetization and film thickness of CoTb layers in (c) and (d) differ from each other, which is taken into account during our calculation of ${\alpha }_{\mathrm{SH}}$.

Figure 4

Comparative measurements of CoTb grown on different spin-orbit materials. (a) and (b) Current-induced switching in $\mathrm{Pt}/\mathrm{CoTb}$ and $\mathrm{Ta}/\mathrm{CoTb}$ samples under positive bias fields. (c) and (d) SOT efficiency $\chi$ vs bias field $H_x$ in $\mathrm{Pt}/\mathrm{CoTb}$ and $\mathrm{Ta}/\mathrm{CoTb}$ samples. Note that compared with the ${\mathrm{Bi}}_2{\mathrm{Se}}_3/\mathrm{CoTb}$ sample, CoTb layers with reduced thickness have been used for Pt and Ta samples, to allow for measurable current-induced switching. (e) Absolute values of the effective spin Hall angles of ${(\mathrm{Bi},\mathrm{Sb})}_2{\mathrm{Te}}_3$, ${\mathrm{Bi}}_2{\mathrm{Se}}_3$, Pt, and Ta measured by our experiments. (f) Normalized power consumption (with Ta set to be unity) for switching FM electrodes in unit magnetic volume using ${(\mathrm{Bi},\mathrm{Sb})}_2{\mathrm{Te}}_3$, ${\mathrm{Bi}}_2{\mathrm{Se}}_3$, Pt, and Ta.