Heat Alone Can Drive Orbital Currents of Electrons, Taiwan Team Reports in First Experimental Evidence
National Taiwan University physicists detect the orbital Nernst effect using a copper oxide, platinum and magnetic-insulator stack, pointing to a way to turn waste heat into spintronic signals.

Physicists at National Taiwan University (NTU) have reported the first experimental evidence that a flow of heat can push electrons' orbital angular momentum sideways through a material, an effect known as the orbital Nernst effect. The study was published in Nature Communications.
The Hall and Nernst effects describe how charge carriers, such as electrons or their spins, develop a transverse response when driven by an electric field or a temperature gradient. The spin Hall and orbital Hall effects have drawn intense interest because an electric current can generate transverse flows of spin and orbital angular momentum. Both forms of angular momentum can be transferred to a material's magnetic moments and, when strong enough, switch its magnetization, a promising way to write information in spintronic memory. Until now, though, the thermal counterpart of the orbital Hall effect had not been shown experimentally. Heat flow is harder to control and characterize than electrical current, and orbital and spin angular momentum are hard to tell apart because they share the same detection geometry.
The NTU team designed a multilayer structure that converts heat-generated orbital angular momentum into a measurable electrical signal. Naturally oxidized copper film (CuOx) served as the source, since earlier work showed it produces substantial orbital currents with negligible spin current. Yttrium iron garnet (YIG), a magnetic insulator, served as the detector. Between them the researchers inserted thin platinum, which translates orbital angular momentum into spin angular momentum through spin-orbit coupling. The spin current then reaches the platinum-YIG interface, where whether it is reflected or absorbed depends on YIG's magnetization direction. The reflected portion becomes a voltage through the inverse spin Hall effect.
Several complications had to be overcome. Wires also develop temperature differences that add a voltage through the charge Seebeck effect, which had to be subtracted. Unintended heat flow perpendicular to the film contaminated the signal with spin Seebeck voltages, and the team separated it by studying how the signals behave when the magnetic field is reversed. Platinum itself produces a spin Nernst signal, so the researchers compared three samples: YIG/Pt, YIG/CuOx and YIG/Pt/CuOx.
The platinum-only sample showed a clear thermal signal, while the copper oxide-only sample showed almost none, as expected if it generates orbital rather than spin momentum. When the two were combined, the measured signal reversed sign, and the estimated spin current reaching YIG rose to more than twice that of the platinum-only sample. That sign reversal and enhancement, the authors say, is compelling transport evidence for an extra heat-driven orbital current in the copper oxide. They estimate its orbital Nernst response is about four times larger than the spin Nernst response of platinum.
The voltages are tiny and serve as detection signals, not usable power. Still, the finding suggests a path for spintronic devices, since heat is unavoidably generated when electronics operate, and converting part of it into angular momentum could give another mechanism for device operation. "The orbital Hall effect has been intensively studied in recent years, while much less is known about its thermal counterpart," said corresponding author Dr. Danru Qu of NTU's Center for Condensed Matter Sciences. "Our work provides experimental evidence that a temperature gradient can generate orbital angular momentum transport." The paper is Li et al., "Transport evidence for the orbital Nernst effect," Nature Communications (2026), DOI 10.1038/s41467-026-77878-2.





