We investigate orbital-to-spin conversion in heavy metals (HMs) by measuring the current-induced damping-like torque in $\mathrm{Ti}/R/{\mathrm{Ni}}_{81}{\mathrm{Fe}}_{19}$ trilayers and $R/{\mathrm{Ni}}_{81}{\mathrm{Fe}}_{19}$ bilayers ($R=\mathrm{Pt}, \alpha \text{−}\mathrm{W}$, or $\beta \text{−}\mathrm{W}$) using spin–torque ferromagnetic resonance. Comparison with the corresponding bilayers reveals a Ti-induced torque that is positive for Pt, negative for $\beta \text{−}\mathrm{W}$, and negligible for $\alpha \text{−}\mathrm{W}$. Analysis with a coupled spin–orbital drift-diffusion model shows that spin- and orbital-sink corrections play only minor roles and that orbital-to-spin conversion in the $\mathrm{HM}$ provides the dominant contribution. The Ti-thickness dependence identifies the bulk orbital Hall effect in Ti as the source of the injected orbital current. The fits yield effective conversion amplitudes of opposite signs and of the same order of magnitude for Pt and $\beta \text{−}\mathrm{W}$, together with nanometer-scale orbital diffusion lengths. The signs of the Pt and $\beta \text{−}\mathrm{W}$ responses follow their respective spin Hall responses, whereas the overall material dependence of the response magnitude does not scale quantitatively with the spin Hall conductivity. The pronounced contrast between the two phases of W shows that the observable coupled spin–orbital transport response is not determined solely by the strength of atomic spin–orbit coupling and highlights the importance of crystal structure.

Close

PrevNext

View figure in article

• Magnetization dynamics
• Spin current
• Spin orbitronics
• Spin-orbit torque