We investigate spin transport by thermally excited spin waves in an antiferromagnetic insulator. Starting from a stochastic Landau-Lifshitz-Gilbert phenomenology, we obtain the out-of-equilibrium spin-wave properties. In linear response to spin biasing and a temperature gradient, we compute the spin transport through a normal-metal-antiferromagnet-normal-metal heterostructure. We show that the spin conductance diverges as one approaches the spin-flop transition; this enhancement of the conductance should be readily observable by sweeping the magnetic field across the spin-flop transition. The results from such experiments may, on the one hand, enhance our understanding of spin transport near a phase transition, and on the other be useful for applications that require a large degree of tunability of spin currents. In contrast, the spin Seebeck coefficient does not diverge at the spin-flop transition. Furthermore, the spin Seebeck coefficient is finite even at zero magnetic field, provided that the normal metal contacts break the symmetry between the antiferromagnetic sublattices.

Antiferromagnets may exhibit spin superfluidity since the dipole interaction is weak. We seek to establish that this phenomenon occurs in insulators such as NiO, which is a good spin conductor according to previous studies. We investigate nonlocal spin transport in a planar antiferromagnetic insulator with a weak uniaxial anisotropy. The anisotropy hinders spin superfluidity by creating a substantial threshold that the current must overcome. Nevertheless, we show that applying a high magnetic field removes this obstacle near the spin-flop transition of the antiferromagnet. Importantly, the spin superfluidity can then persist across many micrometers, even in dirty samples. DOI: 10.1103/PhysRevLett.118.137201 Introduction.-Achieving long-range spin transport is essential in spintronics. In metals, conduction electrons can carry spin information. The spin-diffusion length is generally less than a few hundred nanometers and often as short as a couple of nanometers. However, in ferromagnets there are additional transport channels via spin excitations, typically in the form of spin waves. In ferromagnetic insulators, the absence of noisy itinerant carriers implies less dissipation such that magnons can traverse distances up to several microns [1]. Magnetic low-damping insulators in which new spin transport mechanisms can exist are of interest and can be promising candidates in low-dissipation spintronics.Antiferromagnets (AFMs) have ordered spin configurations, but there is no net magnetization at equilibrium. New observations and advances in our understanding have stimulated increased interest in AFM spintronics [2][3][4][5]. AFMs produce no stray fields that can influence other elements. There are more known high-temperature AFM insulators and semiconductors than their ferromagnetic counterparts. AFMs exhibit transport properties similar to those of ferromagnets. Some of these features are anisotropic magnetoresistance [6], giant magnetoresistance [7], the large anomalous Hall effect in noncollinear AFMs [8], and the spin Hall effect (SHE) [9]. There are also recent investigations of the spin Seebeck effect in AFMs [10][11][12][13]. Additionally, there are observations of spin transport in AFMs via spin pumping from an adjacent ferromagnet into AFMs [14][15][16][17]. In these experiments, it is possible that (evanescent) magnons carry the spin current [18]. A unique aspect of AFMs is that it is possible to trigger ultrafast THz dynamics of the AFM order parameter via charge [19,20] [26][27][28].In this Letter, we investigate spin transport via spin superfluidity (SSF) in AFM insulators. We focus on NiO as

In ferromagnets, magnons may condense into a single quantum state. Analogous to superconductors, this quantum state may support transport without dissipation. Recent works suggest that longitudinal spin transport through a thin-film ferromagnet is an example of spin superfluidity. Although intriguing, this tantalizing picture ignores long-range dipole interactions; we demonstrate that such interactions dramatically affect spin transport. In single-film ferromagnets, "spin superfluidity" only exists at length scales (a few hundred nanometers in yttrium iron garnet) somewhat larger than the exchange length. Over longer distances, dipolar interactions destroy spin superfluidity. Nevertheless, we predict re-emergence of spin superfluidity in tri-layer ferromagnet-normal metal-ferromagnet films of ∼ 1 µm in size. Such systems also exhibit other types of long-range spin transport in samples several micrometers in size.When matter enters a superfluid phase, it behaves like a fluid with zero viscosity and can support currents without dissipation. It has been suggested that certain ferromagnets may exhibit spin superfluidity (SSF) [1][2][3]. The superfluid spin-drag properties induced by spin transfer and spin pumping (SP) in a normal metal-ferromagnetnormal metal system have recently been computed [4][5][6]. Related studies have also explored Josephson spin currents between magnons condensates [7]. Experimental studies have suggested that the temporal decrease of magnon condensates is associated with SSF [8].In the absence of magnetic fields, SSF is indeed an intriguing possibility because its realization would allow spin currents to propagate without significant losses over long distances. These spin transport properties may be useful for low-dissipation interconnects, spin logic devices, and non-volatile magnetic memory devices. Our work demonstrates that SSF can exist in thin-film ferromagnetic systems, but two ferromagnets (rather than one) are required to cancel long-range dipole interactions. We do not observe signatures of long-range SSF in singlefilm ferromagnets.Recent works have hypothesized that easy-plane ferromagnetic thin films exhibit SSF. In such systems, a monotonously precessing magnetization leads to metastable spin-current-carrying states whose topological properties protect against dissipation [3]. Spin relaxation induces a finite resistance proportional to the system size [4]. Nevertheless, ferromagnetic insulators (FIs) have exceptionally low spin dissipation rates, and the spin supercurrent decays over a large length scale. Furthermore, the spin-relaxation-induced algebraic decay of the spin supercurrent significantly differs from the exponential decay of the spin current carried by spin waves [9]. Although magnetic anisotropy destroys the linear SSF response, the spin current is predicted to flow with negligible dissipation when the bias is sufficiently large [4,5].It is well known that long-range dipole interactions dramatically affect the spin-wave dispersion in thin films [10,11]. Low-energy ...

Herein, we study the spin-wave dispersion and dissipation in a ferromagnetic insulator-normal metal-ferromagnetic insulator system. Long-range dynamic coupling because of spin pumping and spin transfer lead to collective magnetic excitations in the two thin-film ferromagnets. In addition, the dynamic dipolar field contributes to the interlayer coupling. By solving the Landau-LifshitzGilbert-Slonczewski equation for macrospin excitations and the exchange-dipole volume as well as surface spin waves, we compute the effect of the dynamic coupling on the resonance frequencies and linewidths of the various modes. The long-wavelength modes may couple acoustically or optically. In the absence of spin-memory loss in the normal metal, the spin-pumping-induced Gilbert damping enhancement of the acoustic mode vanishes, whereas the optical mode acquires a significant Gilbert damping enhancement, comparable to that of a system attached to a perfect spin sink. The dynamic coupling is reduced for short-wavelength spin waves, and there is no synchronization. For intermediate wavelengths, the coupling can be increased by the dipolar field such that the modes in the two ferromagnetic insulators can couple despite possible small frequency asymmetries. The surface waves induced by an easy-axis surface anisotropy exhibit much greater Gilbert damping enhancement. These modes also may acoustically or optically couple, but they are unaffected by thickness asymmetries.

Antiferromagnetic insulators can become active spintronics components by controlling and detecting their dynamics via spin currents in adjacent metals. This cross-talk occurs via spin-transfer and spin-pumping, phenomena that have been predicted to be as strong in antiferromagnets as in ferromagnets. Here, we demonstrate that a temperature gradient drives a significant heat flow from magnons in antiferromagnetic insulators to electrons in adjacent normal metals. The same coefficients as in the spin-transfer and spin-pumping processes also determine the thermal conductance. However, in contrast to ferromagnets, the heat is not transferred via a spin Seebeck effect which is absent in antiferromagnetic insulator-normal metal systems. Instead, the heat is transferred via a large staggered spin Seebeck effect.PACS numbers: 72.25. Mk,72.20.Pa,73.50.Lw,72.10.Di In spintronics, the properties which make antiferromagnets markedly different from ferromagnets also make them attractive in a more dynamic role. Antiferromagnets operate at much higher frequencies and may empower Terahertz circuits. They also have no magnetic stray fields, which therefore enables denser spintronics circuits. For these reasons, antiferromagnets are usually passive spintronics components. However, they can play a role as active components despite their lack of a macroscopic magnetic moment [1-13] and even when they are insulating [10,12,13].We demonstrate that the thermal coupling between antiferromagnetic insulators (AFIs) and normal metals is relatively strong. The strong thermal coupling facilitates several outcomes, can lead to efficient cooling of antiferromagnetic spintronics devices, might function as heat sensors and can reveal valuable information about the high-frequency spin excitations in DC measurements that are complicated to extract with other techniques.Antiferromagnets can produce pure spin currents as large as those produced by ferromagnets. We recently showed that spin pumping may be as operative from antiferromagnets as from ferromagnets [13], in apparent contraction to naive intuition. Furthermore, the efficiency of spin pumping from antiferromagnets to normal metals implies, via Onsager reciprocity relations, that there is a considerable spin-transfer torque on antiferromagnets from a spin accumulation in adjacent normal metals. However, in the absence of external magnetic fields, the spin Seebeck effect vanishes [14]. This fact seems to indicate that spins in antiferromagnets decouple from, or are only weakly connected to, heat currents and temperature gradients in adjacent normal metals.To the contrary, we find that the thermal coupling constant is orders of magnitude stronger than its ferromagnetic counterpart. This radical difference is caused by the large exchange field in antiferromagnets that governs the heat transfer rather than the much smaller anisotropy fields or external magnetic fields in ferromagnets. The thermal coupling between antiferromagnetic insulators and normal metals is associated with a staggered s...