Spin transport by measurement and inference

Quantifying spin transport in thin films

A thin native oxide on permalloy (Py) inflates FMR linewidths and biases the Gilbert damping parameter α upward, invisible unless you track the slope across many frequencies and cross-check surface chemistry by XPS. Extracting the spin Hall angle of the adjacent heavy-metal layer via the inverse spin Hall effect (iSHE) is further complicated by anisotropic magnetoresistance (AMR) of the ferromagnet, which generates a voltage at the same resonance condition that cannot be separated from the iSHE signal by geometry alone.

We measured four bilayer systems (Py/Pt, Py/W, Py/Cu, Py/Ti) and later YIG/Pt and YIG/W. For α, lock-in detection with field modulation extracted derivative-Lorentzian line shapes, and XPS-guided capping-layer selection kept the oxide contribution identifiable. For iSHE, Py/Ti served as the AMR reference: titanium has negligible spin–orbit coupling, so the voltage in that stack is pure baseline. Subtracting it from Py/Pt and Py/W isolates the symmetric iSHE component. Switching to YIG (an insulating ferromagnet with no charge current in the magnetic layer) removed the leakage entirely, leaving purely Lorentzian iSHE line shapes.

Platinum’s spin-orbit coupling was the strongest by both measures simultaneously: largest damping enhancement and largest iSHE amplitude, a consistency that points to a real material effect rather than artefact. Tungsten was clearly second; copper and titanium were effectively zero, consistent with their weak spin–orbit coupling. The sign reversal between YIG/Pt and YIG/W confirmed the theoretical prediction for the spin Hall angle. To keep the FMR rig running through overnight sweeps, we added a low-cost IoT monitor (ESP8266 + Hall flow sensor) with automated email alerts; it caught at least two coolant interruptions before they could spoil measurements.

Methods and data: Full report (PDF)