We theoretically and numerically study the terahertz auto-oscillations, or self-oscillations, in thin-film metallic noncollinear coplanar antiferromagnets (AFMs), such as ${\mathrm{Mn}}_{3}\mathrm{Sn}$ and ${\mathrm{Mn}}_{3}\mathrm{Ir}$, under the effect of antidamping spin torque with spin polarization perpendicular to the plane of the film. To obtain the order parameter dynamics in these AFMs, we solve three Landau-Lifshitz-Gilbert equations coupled by exchange interactions assuming both single- and multidomain (micromagnetics) dynamical processes. In the limit of a strong exchange interaction, the oscillatory dynamics of the order parameter in these AFMs, which have opposite chiralities, could be mapped to that of two damped-driven pendulums with significant differences in the magnitude of the threshold currents and the range of frequency of operation. The theoretical framework allows us to identify the input current requirements as a function of the material and geometry parameters for exciting an oscillatory response. We also obtain a closed-form approximate solution of the oscillation frequency for large input currents in the case of both ${\mathrm{Mn}}_{3}\mathrm{Ir}$ and ${\mathrm{Mn}}_{3}\mathrm{Sn}$. Our analytical predictions of threshold current and oscillation frequency agree well with the numerical results and thus can be used as compact models to design and optimize the auto-oscillator. Employing a circuit model, based on the principle of tunnel anisotropy magnetoresistance, we present detailed models of the output power and efficiency versus oscillation frequency of the auto-oscillator. Finally, we explore the spiking dynamics of two unidirectional as well as bidirectional coupled AFM oscillators using nonlinear damped-driven pendulum equations. Our results could be a starting point for building experimental setups to demonstrate auto-oscillations in metallic AFMs, which have potential applications in terahertz sensing, imaging, and neuromorphic computing based on oscillatory or spiking neurons.
Read full abstract