Cornell University, Ithaca, New York 14850, USA.
Kavli Institute at Cornell, Ithaca, New York 14853, USA.
Phys Rev Lett. 2019 Aug 2;123(5):057203. doi: 10.1103/PhysRevLett.123.057203.
The effective spin-mixing conductance (G_{eff}^{↑↓}) of a heavy-metal-ferromagnet (HM-FM) interface characterizes the efficiency of the interfacial spin transport. Accurately determining G_{eff}^{↑↓} is critical to the quantitative understanding of measurements of direct and inverse spin Hall effects. G_{eff}^{↑↓} is typically ascertained from the inverse dependence of magnetic damping on the FM thickness under the assumption that spin pumping is the dominant mechanism affecting this dependence. We report that this assumption fails badly in many in-plane magnetized prototypical HM-FM systems in the nanometer-scale thickness regime. Instead, the majority of the damping is from two-magnon scattering at the FM interface, while spin-memory-loss scattering at the interface can also be significant. If these two effects are neglected, the results will be an unphysical "giant" apparent G_{eff}^{↑↓} and hence considerable underestimation of both the spin Hall ratio and the spin Hall conductivity in inverse or direct spin Hall experiments.
重金属-铁磁体(HM-FM)界面的有效自旋混合电导(G_{eff}^{↑↓})表征了界面自旋输运的效率。准确确定 G_{eff}^{↑↓}对于定量理解直接和反向自旋霍尔效应的测量至关重要。G_{eff}^{↑↓}通常是从假设自旋泵浦是影响这种依赖性的主要机制的情况下,从 FM 厚度的磁阻尼的逆依赖性来确定的。我们报告说,在许多纳米级厚度范围内的平面磁化原型 HM-FM 系统中,这种假设严重失败。相反,大部分阻尼来自 FM 界面处的双磁子散射,而界面处的自旋记忆损耗散射也可能很重要。如果忽略这两个效应,结果将是不真实的“巨大”表观 G_{eff}^{↑↓},因此在反向或正向自旋霍尔实验中,对自旋霍尔比和自旋霍尔电导率的估计都会大大低估。