Nanoscale Mapping of Magnetic Auto-Oscillations with a Single Spin Sensor

Spin Hall nano-oscillators convert DC to magnetic auto-oscillations in the microwave regime. Current research on these and similar devices is dedicated to creating next-generation energy-efficient hardware for communication technologies. Despite intensive research on magnetic auto-oscillations within the past decade, the nanoscale mapping of those dynamics remained a challenge. In consequence, insight could mainly be gained via micromagnetic simulations. Scientists from Max Planck Institute of Solid State Research in collaboration with University of Stuttgart have shown a first direct experimental measurement of freerunning auto-oscillation in such an oscillator device using the high sensitivity and spatial resolution of a color-center based quantum sensing approach.

Magnetic auto-oscillations can be understood as the damping-compensated precession of the magnetization around the equilibrium axis given by the magnetic field. Due to this damping-compensation, such oscillations persist at a stable precession angle and frequency.
Naturally, magnetic oscillations face damping which results in a loss of of amplitude over time. However, this damping can be compensated via injection of a spin current into the ferromagnetic material. In a spin Hall nano-oscillator, the spin Hall effect is used to generate a pure spin current which is the transport of spins without net transport of charges. Fortunately, this can be simply done by depositing two metals about each other and applying DC (Fig. 1). In the present case Platinum is used to generated a sufficient spin Hall effect and a NiFe alloy is used as ferromagnetic material with low damping.

Fig. 1: Cross section of the spin Hall nano-oscillator. The DC is converted in a pure spin current within Pt via spin Hall effect. The spin-polarized electrons diffuse in to the ferromagnetic NiFe layer. The transfer of angular momentum on the magnetization compensates the damping leading to auto-oscillations of the magnetization. This process starts preferred at areas of low internalmagnetic field (lower damping). The auto-oscillations generate a microwave field in close vicinity which interacts with the spin sensor in thediamond tip. — Fig. 1: Cross section of the spin Hall nano-oscillator. The DC is converted in a pure spin current within Pt via spin Hall effect. The spin-polarized electrons diffuse in to the ferromagnetic NiFe layer. The transfer of angular momentum on the magnetization compensates the damping leading to auto-oscillations of the magnetization. This process starts preferred at areas of low internal magnetic field (lower damping). The auto-oscillations generate a microwave field in close vicinity which interacts with the spin sensor in the diamond tip.

Fig. 1: Cross section of the spin Hall nano-oscillator. The DC is converted in a pure spin current within Pt via spin Hall effect. The spin-polarized electrons diffuse in to the ferromagnetic NiFe layer. The transfer of angular momentum on the magnetization compensates the damping leading to auto-oscillations of the magnetization. This process starts preferred at areas of low internal
magnetic field (lower damping). The auto-oscillations generate a microwave field in close vicinity which interacts with the spin sensor in the
diamond tip.

In order to generate a sufficient DC density to overcome the damping, the bilayer is structured in wires with well defined constrictions where the DC density maximum is reached. Interestingly, the DC density distribution is inhomogeneous with the maxima at the constriction edges.
Beside this spatially inhomogeneous antidamping effect, the damping is spatially inhomogeneous as well since it is linked to the distribution of the internal magnetic field which is varied due to the shape anisotropy of the device structure.

Moreover, the magnetization is affected by the Oersted field and Joule heating generated by the applied DC. Furthermore, auto-oscillations are strongly nonlinear dynamics, since the change of precession angle results in a variation of the effective magnetic field in which the magnetization is placed. This forms a feedback loop which couples precession amplitude to the auto-oscillation frequency.
In summary, the generation of auto-oscillations is a sophisticated interplay between several spatially inhomogeneous effects which depend on the device structure, materials, magnetic field and applied DC. The prediction via micromagnetic simulations is therefore limited. Highly sensitive and spatially-resolved mapping of auto-oscillation modes is imperative to understand and optimize such devices for applications in communication technologies. For the first time it has been possible to map the distribution of these auto-oscillations and explain their localization using a nitrogen-vacancy center as a quantum sensor in a diamond tip as a scanning-probe. These measurements revealed for the first time that auto-oscillations are localized at magnetic field minima.
This localization can be explained by the spinwave band gap present at low frequencies at these conditions. In the surrounding area, there are simply no spin-wave states at the auto-oscillation frequency and, therefore, spin-wave propagation is impossible. As a consequence, the magnetic field minima do not only determine the position of auto-oscillations but also act as spin-wave resonators.
These results confirm previous results from micromagnetic simulations and enable experimental investigation of auto-oscillation generation in real devices affected by defects from to the fabrication process. This paves the way for the next generation of magnetic nano-oscillator devices which are highly attractive as nanoscale microwave voltage and spin wave sources, as well as for neuromorphic computing hardware.