Event Details

The flux of heat carried by magnons in the presence of a temperature gradient corresponds to a flux of magnetization, which in turn can interact with conduction electrons. In the spin-Seebeck effect (SSE), the interaction takes place across an interface between a ferromagnet (typically YIG) and a normal metal (typically Pt), and is parametrized by the spin mixing conductance. The free electrons in the Pt are then spin-polarized over a spin diffusion length, and give rise to a transverse voltage via the inverse spin-Hall effect in the Pt. The loss mechanisms are primarily the inefficiency of the inverse spin Hall conversion of a magnon to a charge flux, and the limitations imposed by the magnon thermal length near the Pt/YIG interface. In 2011, Lucassen [1] suggested that magnon-electron drag is another form of advective thermal spin/charge transport akin to the SSE, but in a uniform material, eliminating the need for an interface. In the colloquium we provided proof of this in the thermopower, but not for the Nernst coefficient. Indeed, it is known that phonon-drag cannot per se generate a skew force and thus a Nernst effect [2], and therefore this is likely the case for magnon-drag as well. In contrast, here we present a new model based on ambipolar transport. Spin-up and spin-down electrons in Fe are considered as charge carriers with separate magnon-drag Seebeck coefficients. The difference between these partial Seebeck coefficients leads to a large magnon-drag Nernst coefficient even in the absence of a skew force. Here we will provide experimental data and a new tentative model for magnon-drag Nernst coefficients. We then explore thermal magnon transport in a uniaxial antiferromagnet KNiF3 that promises to become an experimental platform for thermally induced spin transport and possibly dynamically-induced spin superfluidity [3]. The existence of a large number of insulating and semiconducting antiferromagnets in which advective thermal spin/charge transport is possible broadens the range of possible materials for this research.

[1] Lucassen et al., Appl. Phys. Lett. 99, 262506 (2011);
[2] Korenblit, L. Y., Sov. Phys. Semiconductors2, 1192 (1962);
[3] Takei et al, Phys. Rev. B 90, 094408 (2014)