Strong voltage-induced tunability of threshold current and frequency in spin Hall nano-oscillators

Spin Hall nano-oscillators (SHNOs) utilize pure spin currents to excite auto-oscillations at microwave frequencies in nanoscopic regions of magnetic thin films^1-3. Thanks to their wide frequency tunability^1,2, their robust mutual synchronization in chains as well as in 2D arrays^1,3, and their CMOS compatibility², SHNOs have recently emerged as a promising spintronic device for neuromorphic computing³. While the SHNO frequency can be tuned with both current and external field, the lack of fine individual SHNO control in synchronized networks limits cognitive tasks in oscillator-based computing. It would be highly advantageous if an energy efficient route such as electrostatic gating can independently tune the SHNO characteristics in a synchronized network to perform more complex neuromorphic tasks^3,4.

Here we demonstrate how electrostatic gating in nano-constriction based W(5nm)/CoFeB(1.7nm)/MgO(2nm) SHNOs can lead to substantial voltage-controlled tunability of threshold current and auto-oscillation frequency. High frequency microwave measurements show a large overall modulation of about 22% in threshold current with △V_G= 4V and a moderate frequency tunability of 12 MHz/V. Our detailed analysis based on ST-FMR measurements combined with micromagnetic simulations unveil that the observed tunabilities are caused by minor voltage-induced changes in the perpendicular magnetic anisotropy, which, in our case, not only tunes the frequency but also significantly modifies the localization of the auto-oscillating mode resulting in a large change of the effective damping. Our demonstration introduces a new insight to critically control the effective damping and represents a significant step towards the realization of complex neuromorphic tasks^3,4.

References
¹ A. A. Awad, et al., Nat. Phys. 13, 292-299 (2017).
² H. Fulara, et al., Sci. Adv. 5, eaax8467 (2019).
³ M. Zahedinejad, et al., arXiv:1812.09630 (2018); Nat. Nanotech., accepted.
⁴ M. Romera, et al., Nature 563, 230 (2018).