2025/07/14 by Utkarsh Shashank, Shashank, Utkarsh, Akash Kumar +17 · 1 citation
Engineering · Physics and Astronomy · #Advanced Materials Characterization Techniques #Applied Physics (physics.app-ph) #FOS: Physical sciences #Magnetic properties of thin films #Mesoscale and Nanoscale Physics (cond-mat.mes-hall) #Quantum and electron transport phenomena
paper · pdf · doi:10.48550/arxiv.2507.10219
openalex publication_date 2025/07/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Spin-orbit-torque-driven auto-oscillations in spin Hall nano-oscillators (SHNOs) offer a transformative pathway toward energy-efficient, nanoscale microwave devices for next-generation neuromorphic computing and high-frequency technologies. A key requirement for achieving robust, sustained oscillations is reducing the threshold current (Ith), strongly governed by spin Hall efficiency (θSH). However, conventional strategies to enhance θSH face trade-offs, including high longitudinal resistivity, interfacial effects, and symmetry-breaking torques that limit performance. Here, we demonstrate a substantial enhancement of the bulk spin Hall effect in PtBi alloys, achieving over a threefold increase in θSH, from 0.07 in pure Pt to 0.24 in Pt94.0Bi6.0 and 0.19 in Pt91.3Bi8.7, as extracted from DC-bias spin-torque ferromagnetic resonance. The enhanced θSH originates from bulk-dominated, extrinsic side-jump scattering across all PtBi compositions. Correspondingly, we observe a 42% and 32% reduction in Ith in 100 nm SHNOs based on Co40Fe40B20(3 nm)/Pt94.0Bi6.0(4 nm) and Co40Fe40B20(3 nm)/Pt91.3Bi8.7(4 nm), respectively. Structural characterization reveals reduced Pt crystallinity, along with emergence of preferred crystallographic orientations upon introducing higher Bi concentrations. Together, these results position PtBi alloys as a compelling alternative to conventional 5d transition metals, enabling enhanced θSH and significantly lower Ith, thus opening new avenues for energy-efficient neuromorphic computing and magnetic random access memory.