Phys. Rev. Lett. 111, 176601 (2013)

Phys. Rev. Lett. 111, 187201 (2013)

In the spin Seebeck effect, a temperature gradient applied to a ferromagnetic/non-magnetic bilayer gives rise to a spin current — currents of angular momentum with zero net charge transport — that can be detected by the inverse spin Hall effect in the non-magnetic layer. The spin Seebeck effect has attracted attention because of the possibility of controlling spin currents by thermal flow; however, there is an ongoing debate over the origin and magnitude of the effect due to the potential contamination of the measured spin Seebeck signal by other effects, namely the anomalous Nernst effect and anisotropic magnetoresistance. Two independent research teams have now tried to address this issue and report measurements in yttrium iron garnet/Pt thin-film bilayers and permalloy thin films grown on different substrates.

The two groups use different measurement configurations, longitudinal and transverse, in which a spin current is generated by the spin Seebeck effect either parallel or perpendicular to the thermal gradient, respectively. Using a longitudinal configuration, Sebastian Gönnenwein at the Walther-Meißner-Institut and colleagues in Germany, The Netherlands, China and Japan measure spin pumping, spin Seebeck effect and spin Hall magnetoresistance in a series of 20 yttrium iron garnet/Pt samples, and model these effects with the same set of spin-dependent parameters, demonstrating their spintronic origin. Alternatively, and using a transverse configuration, Christian Back and colleagues at the University of Regensburg and Bielefeld University study the effect of different substrates on the magnitude of the spin Seebeck effect, and find that, regardless of which substrate is used, the contribution of the spin Seebeck effect is negligible compared with those of the anomalous Nernst effect and anisotropic magnetoresistance, in contrast to previous results.