Unveiling Interfacial Spin-Orbit Torques in the Atomically Thin Limit: the Key Role of Skew Scattering
Aires Ferreira, University of York
July 28, 2020
Fueled by the discovery of ferromagnetism in 2D crystals in 2017, recent works have achieved spin-orbit torque (SOT) switching of ultra-thin van-der-Waals-bonded ferromagnets approaching the 2D limit [1]. Conversely, nonmagnetic 2D crystals can be employed as an efficient source of nonequilibrium spin polarization [2]. Experiments with atomically thin WTe2 demonstrated sought-after anti-damping SOT (ASOT) required for switching of perpendicular magnetization [3]. While these early findings represent a stepping-stone towards envisaged all-electrical spin memories built entirely from 2D crystals [4], the driving force behind purely interfacial ASOT remains generally poorly understood.
In this talk I will argue that left-right scattering asymmetry plays a key, albeit hitherto neglected role in interfacial SOT. Starting from a unified theory of 2D crystal-ferromagnet interfaces, I will show that skew scattering generates a number of measurable effects (recently predicted in [5]), including: (i) a current-induced collinear spin polarization leading to sizeable ASOT and (ii) a non-equilibrium out-of-plane spin polarization which can be manipulated by controlling the current direction. Interestingly, the proposed Fermi surface mechanism is sensitive to sublattice-symmetry breaking and presence of impurity states near the Fermi level, resulting in a rich SOT phenomenology. Our accurate microscopic calculations indicate that the electrical torque efficiency can be tuned by up to 2 orders of magnitude by engineering the disorder landscape (e.g. by introduction of resonant impurities) [5]. The skew scattering mechanism is operative in realistic devices and is expected to dominate the ASOT angular dependence in weakly disordered interfaces. Finally, I will briefly discuss the implications of these theoretical findings for our general understanding of emergent SOTs at heterointerfaces [6].
[1] M. Alghamdi et al., Nano Lett. 19, 4400 (2019). X. Wang et al., Science Adv. 5, eaaw8904 (2019).
[2] M. Offidani, M. Milletari, R. Raimondi & A. Ferreira, Phys. Rev. Lett. 119, 196801 (2018).
[3] D. MacNeill et al., Nature Phys.13, 300 (2017).
[4] K. Dolui et al., Nano Lett. 20, 2288 (2020).
[5] F. Sousa, G. Tatara & A. Ferreira, arXiv:2005.09670 (2020).
[6] A. Manchon et al., Rev. Mod. Phys. 91, 035004 (2019).