Proceedings Article | 18 September 2018
Tom Silva, Eric R. Edwards, Hans Nembach, Olof Karis, Mathias Weiler, Andrew Berger
KEYWORDS: Platinum, Diffusion, Interfaces, Ferromagnetics, Magnetism, Metals, Spectroscopy, Thin films, Data conversion, Transparency
Functional spintronic devices rely on spin-charge interconversion effects, such as the reciprocal processes of electric field-driven spin torque and magnetization dynamics-driven spin and charge flow. Both damping-like and field-like spin-orbit torques have been observed in the forward process of current-driven spin torque and damping-like inverse spin-orbit torque has been well-studied via spin pumping into heavy metal layers. Here we demonstrate that established microwave transmission spectroscopy of ferromagnet/normal metal bilayers under ferromagnetic resonance can be used to inductively detect the AC charge currents driven by the inverse spin-charge conversion processes. This technique relies on vector network analyzer ferromagnetic resonance (VNA-FMR) measurements. We show that in addition to the commonly-extracted spectroscopic information, VNA-FMR measurements can be used to quantify the magnitude and phase of all AC charge currents in the sample, including those due to spin pumping and spin-charge conversion. Our findings reveal that Permalloy/Pt bilayers exhibit both damping-like and field-like inverse spin-orbit torques. While the magnitudes of both the damping-like and field-like inverse spin-orbit torque are of comparable scale to prior reported values for similar material systems, we observed a significant dependence of the damping-like magnitude on the order of deposition. This suggests interface quality plays an important role in the overall strength of the damping-like spin-to-charge conversion.
Spin memory loss (SML) [1] and proximity-induced magnetic moments at the FM/NM interface [2] have been invoked to explain the large damping enhancement caused by thin NM films even when the NM thickness is less than its spin diffusion length. In this model, spin loss at the FM/NM interface acts as an additional parallel spin relaxation pathway to that of spin pumping and diffusion into the Pt bulk. From damping measurements alone, the relative contributions of these mechanisms are not resolvable. In this work, we show that a self-consistent fit of Gilbert damping and damping-like iSOT versus Pt thickness—where both sets of data are described by the same spin diffusion length—makes it possible to separate these sources of damping. Furthermore, this data analysis methodology allows for unambiguous determination of the spin-mixing conductance at the FM/NM interface. We therefore can determine the spin Hall conductivity (or spin Hall angle) without having to refer to spin transport parameters, e.g. the spin-mixing conductance and spin diffusion length, as determined from measurements performed on dissimilar samples or theoretical idealized values. For our samples of Pt deposited on Permalloy, only 37 ± 6% of the total damping enhancement from the Pt film is attributable to spin pumping into the Pt layer when the Pt thickness is much greater than the spin diffusion length. The self-consistent fit also results in a spin diffusion of length of (4.2 ± 0.1) nm, and a spin mixing conductance of (130,000 ± 20,000) 1/(μΩ cm^2), which is in good agreement with the maximum theoretical value for Pt of 107,000 1/(μΩ cm^2) [3], given the estimated error, and σ_SH = (2.36 ± 0.04) 1/(μΩ m). This corresponds to a spin Hall angle of 0.387 ± 0.008. While this θ_SH is among the largest reported for Pt [4, 5], it is a necessary logical conclusion that with less spin current driven into the NM (on account of SML), a larger spin-to-charge conversion efficiency is required to fit the data than would be otherwise obtained if the SML were negligible. We furthermore stress that the phenomenological value for the damping-like spin orbit torque is comparable to that measured with other techniques [5-7]. This indicates that the Pt layer in our samples behaves conventionally, and stresses the importance of characterizing spin loss mechanisms to optimize SOT for magnetic switching.
REFERENCES
[1] J.-C. Rojas-Sánchez, N. Reyren, P. Laczkowski, et al., "Spin Pumping and Inverse Spin Hall Effect in Platinum: The Essential Role of Spin-Memory Loss at Metallic Interfaces," Phys. Rev. Lett., vol. 112, p. 106602 (2014).
[2] M. Caminale, A. Ghosh, S. Auffret, et al., "Spin pumping damping and magnetic proximity effect in Pd and Pt spin-sink layers," Physical Review B, vol. 94, p. 014414 (2016).
[3] Y. Liu, Z. Yuan, R. J. H. Wesselink, et al., "Interface Enhancement of Gilbert Damping from First Principles," Physical Review Letters, vol. 113, p. 207202 (2014).
[4] W. Zhang, W. Han, X. Jiang, et al., "Role of transparency of platinum-ferromagnet interfaces in determining the intrinsic magnitude of the spin Hall effect," Nat Phys, Article vol. 11, pp. 496-502 (2015).
[5] C.-F. Pai, Y. Ou, L. H. Vilela-Leão, et al., "Dependence of the efficiency of spin Hall torque on the transparency of Pt/ferromagnetic layer interfaces," Phys. Rev. B, vol. 92, p. 064426 (2015).
[6] K. Garello, I. M. Miron, C. O. Avci, et al., "Symmetry and magnitude of spin-orbit torques in ferromagnetic heterostructures," Nat. Nano., vol. 8, pp. 587-593 (2013).
[7] M.-H. Nguyen, D. C. Ralph, and R. A. Buhrman, "Spin Torque Study of the Spin Hall Conductivity and Spin Diffusion Length in Platinum Thin Films with Varying Resistivity," Physical Review Letters, vol. 116, p. 126601 (2016).
