Experimental Test of the Spin Mixing Interface Conductivity Concept

Abstract

We perform a quantitative, comparative study of the spin pumping, spin Seebeck and spin Hall magnetoresistance effects, all detected via the inverse spin Hall effect in a series of over 20 yttrium iron garnet/Pt samples. Our experimental results fully support present, exclusively spin currentbased, theoretical models using a single set of plausible parameters for spin mixing conductance, spin Hall angle and spin diffusion length. Our findings establish the purely spintronic nature of the aforementioned effects and provide a quantitative description in particular of the spin Seebeck effect.Pure spin currents present a new paradigm in spintronics [1, 2] and spin caloritronics [3]. In particular, spin currents are the origin of spin pumping [4,5], the spin Seebeck effect [6,7] and the spin Hall magnetoresistance (SMR) [8][9][10]. Taken alone, all these effects have been extensively studied, both experimentally [6-9, 11-13] and theoretically [4,[14][15][16][17][18]. From a theoretical point of view, all these effects are governed by the generation of a current of angular momentum via a non-equilibrium process. The flow of this spin current across a ferromagnet/normal metal interface can then be detected. The relevant interface property that determines the spin current transport thereby is the spin mixing conductance. Nevertheless, there has been an ongoing debate regarding the physical origin of the measurement data acquired in spin Seebeck and SMR experiments due to possible contamination with anomalous Nernst effect [19][20][21] or anisotropic magnetoresistance [22,23] caused by static proximity polarization of the normal metal [23]. To settle this issue, a rigorous check of the consistency of the spin-current based physical models across all three effects is needed. If possible contamination effects are absent, according to the spin mixing conductance concept [24], there should exist a generalized Ohm's law between the interfacial spin current and the energy associated with the corresponding non-equilibrium process. This relation should invariably hold for the spin pumping, spin Seebeck and spin Hall magnetoresistance effects, as they are all based on the generation and detection of interfacial, nonequilibrium spin currents. We here put forward heuristic arguments that are strongly supported by experimental evidence for a scaling law that links all aforementioned spin(calori)tronic effects on a fundamental level and allows to trace back their origin to pure spin currents. (c) The spin Hall magnetoresistance is due to the torque exerted on M by an appropriately polarized Js which yields a change in the reflected spin current J r s . The interconversion between Js (J r s ) and the charge currents Jc (J r c ) are due to the (inverse) spin Hall effect in the normal metal.[schematically depicted in Fig. 1(a)], we place YIG / Pt bilayers in a microwave cavity operated at ν = 9.85 GHz to resonantly excite magnetization dynamics. The emission of a spin current density J s across the bilayer interface into the Pt provides...

Current-induced spin torque resonance of a magnetic insulator

“…Moreover, most parameters are constrained (by literature or other measurements), leaving only charge current density J c and intrinsic damping parameter α 0 as free parameters in the model calculations. These, however, can also be estimated well from the applied microwave source power, impedance matching characteristics, and the damping parameters extracted from spin pumping experiments on our YIG/Pt samples [32]. The spin transport parameters are spin diffusion length λ = 1.5 nm, spin Hall angle θ SH = 0.11, and spin mixing conductance Re(G ↑↓ ) = 4 × 10 14 −1 m −2 , which have proven to be robust to variations of other sample parameters (Refs.…”

“…The spin transport parameters are spin diffusion length λ = 1.5 nm, spin Hall angle θ SH = 0.11, and spin mixing conductance Re(G ↑↓ ) = 4 × 10 14 −1 m −2 , which have proven to be robust to variations of other sample parameters (Refs. [30,32]). The phase δ = −63 • is inferred from additional measurements with magnetic fields oriented at a slight angle to the film normal (see Sec.…”

Modulation of pure spin currents with a ferromagnetic insulator

“…One could take advantage of the spin-mixing conductance concept [5,6] at nonmagnetic metal (NM)/ferromagnetic insulator (FMI) interfaces, which governs the interaction between the spin currents present at the NM and the magnetization of the FMI. This concept is the basis of new spindependent phenomena, including spin pumping [6][7][8][9][10][11][12], spin Seebeck effect [6,13], and spin Hall magnetoresistance (SMR) [6,[14][15][16][17][18]. In these cases, a NM with large spin-orbit coupling is required to convert the involved spin currents into charge currents via the inverse spin Hall effect [19].…”

“…Such effects are unlikely in Cu/YIG. It is worth noting that the G r of a NM/YIG interface, for a NM with a negligible spin-orbit coupling, was not experimentally measured before due to the need of the inverse spin Hall effect (and thus a high spin-orbit coupling metal) in the experiments made so far [6][7][8][9][10][11][12][13][14][15][16].…”

Large-Pore Mesoporous Ho3Fe5O12 Thin Films with a Strong Room-Temperature Perpendicular Magnetic Anisotropy by Sol–Gel Processing

“…[4][5][6] Among them are solution-processing routes such as solvothermal and sol-gel methods, physical deposition techniques, mechanochemical approaches, and others. [7][8][9][10][11] As for thin films, materials with a perpendicular magnetic anisotropy have been the subject of significant interest over the past decade due to their potential for application in nanomagnetics [12][13][14] (e.g., as storage/recording media) and spintronics [15][16][17] (e.g., as spin valves). Nevertheless, they are also highly interesting from the viewpoint of basic research.…”