
Physics
Magnetic Force Microscopy of Operational Spin Hall Nano-oscillators
S. Banuazizi, M. Kargar, et al.
Explore groundbreaking research on a novel method for analyzing magnetic nanodevices using magnetic force microscopy, as conducted by S.A.H. Banuazizi, M. Kargar, A.A. Awad, and A. Dmitriev. This innovative approach sheds light on the operational profiles of spin Hall nano-oscillators, revealing fascinating insights into current density and induced Oersted fields.
~3 min • Beginner • English
Introduction
The study investigates how the Oersted field generated by drive current in spin Hall nano-oscillators (SHNOs) modifies the effective magnetic field landscape and impacts key emission characteristics. SHNOs, based on spin currents from the spin Hall effect exerting spin-transfer torque on an adjacent ferromagnet, are easy to fabricate and exhibit promising high-frequency emission properties. A clear understanding and direct measurement of the magnetization properties of SHNOs in their operational state is crucial. The authors propose and demonstrate that magnetic force microscopy (MFM) can be adapted to probe the spatial profile of operational SHNOs despite the mechanical and spatial constraints of MFM setups.
Literature Review
The paper positions SHNOs within prior work on spin-torque-driven auto-oscillators and highlights their dependence on external magnetic field and DC current. It emphasizes that the Oersted field induced by the DC current is a significant contributor to the effective magnetic field landscape in SHNOs, affecting oscillation frequencies and spin-wave modes. Prior MFM approaches for magnetic imaging are referenced, motivating the adaptation of MFM to operational devices, but the paper focuses primarily on methodology rather than an extensive literature survey.
Methodology
Device fabrication: Nanoconstriction-based SHNOs were fabricated as NiFe (permalloy, Ni80Fe20; 5 nm)/Pt (6 nm) bilayers magnetron sputtered at room temperature onto sapphire C-plane substrates (18 x 18 mm) with an in-situ 5 nm SiO2 capping layer. The base pressure during deposition was <3 x 10^-? Torr (value partially truncated in the excerpt). The layers were patterned into 4 x 12 µm rectangles with two ~50 nm tip-radius indentations to form nanoconstrictions with nominal widths dNC = 80–300 nm via e-beam lithography and argon ion milling using a negative e-beam resist as the mask. Ultrasmall constrictions were achieved by ion-beam milling at 45° to exploit lateral erosion of the mask. Electrical access was provided by long coplanar waveguides fabricated by optical lithography and lift-off (980 nm Cu/20 nm Au), designed longer than conventional to fit under the MFM head and allow stable microwave probing.
Measurement setup: An Asylum Research MFP-3D-SA MFM system was extended with a custom microwave probe station. A nonmagnetic GSG microwave probe (GGB, customized with an extended coax) was mounted on an XYZ micromanipulator fixed to the MFM stage’s slider to achieve low-profile access under the MFM head. A variable field module (VFM2) with permanent magnets applied uniform in-plane fields (up to at least ±0.8 T with ~1 G resolution). A microscope camera assisted precise probe-sample alignment. A high-frequency bias-T enabled DC current injection and RF extraction; RF signals were amplified by a low-noise amplifier and recorded by a spectrum analyzer (RBW 1 MHz). For MFM, a standard OMCL-AC240TS cantilever with MFM coating was used; a customized high-resolution MFM probe (per ref. 42) with ~1 µm length, 150 nm diameter, cone-shaped head was also fabricated for improved spatial resolution (~10 nm stated in simulations). The chip containing 20 SHNOs (with long waveguides) was mounted on VFM2, centered between magnet poles for field uniformity; the chip could be rotated to set the in-plane field angle (e.g., 24°).
Simulations: 3D finite-element simulations (COMSOL Multiphysics) modeled current flow and the induced Oersted field in the NiFe/Pt SHNOs (example for dNC = 300 nm). Simulated the y-component of current density and the x-component of the Oersted field (H_Oe) at the top of the Pt under a DC current (e.g., Idc = 6 mA), revealing a central region with elevated H_Oe.
Experimental conditions: SHNOs with dNC from 80 to 300 nm were studied under in-plane magnetic fields and DC currents sufficient to induce auto-oscillations. Representative spectra are shown for dNC = 150 nm at I (reported as Iac) = 2.7 mA, H = 500 Oe, in-plane angle = 24°. MFM scans were performed on dNC = 300 nm devices under H = 800 Oe, Idc = -6 mA; H = 800 Oe, Idc = +6 mA; and H = 1600 Oe, Idc = +6 mA; all at in-plane angle 24°.
Data acquisition: Output RF power spectral density was recorded while scanning conditions were established for MFM imaging. AFM topography and MFM phase shift images (lift mode) were collected over the same areas to correlate structural and magnetic contrasts.
Key Findings
- Simulations: COMSOL results revealed a strong spatial redistribution of current density in the nanoconstriction, producing a pronounced x-component of H_Oe with a higher magnitude in a circular region centered on the constriction (example shown at Idc = 6 mA for dNC = 300 nm).
- Microwave spectroscopy: Operational SHNOs exhibited auto-oscillation signals; a representative PSD showed emission in the 3.6–4.0 GHz range for a device with dNC = 150 nm at reported I = 2.7 mA, H = 500 Oe, angle = 24°. Auto-oscillations were detected across tens of devices with reproducible results.
- MFM imaging: Under applied in-plane magnetic fields and DC bias, MFM phase images of dNC = 300 nm devices showed dark contrast regions at the constriction consistent with areas of high current density and elevated H_Oe predicted by simulations. No such feature appeared without an applied magnetic field, indicating the field-dependent nature of the observed contrast.
- Temporal resolution constraint: The MFM tapping frequency (kHz) is far below the SHNO emission frequencies (3–9 GHz), precluding direct detection of individual magnetization oscillations with the present system; nevertheless, the spatial profile associated with the DC-current-induced Oersted field can be imaged during operation.
Discussion
The work demonstrates that MFM, when integrated with a custom microwave probe station and variable in-plane magnetic fields, can map the spatial profile of an operational SHNO, specifically revealing contrast correlated with the DC-current-induced Oersted field near the nanoconstriction. This addresses the need for direct spatially resolved measurements of the effective field landscape in working SHNOs, which is crucial for understanding and optimizing their emission characteristics. The agreement between simulated H_Oe distributions and MFM phase contrast supports the interpretation that the observed features stem from current-induced fields rather than direct imaging of GHz magnetization oscillations (which are beyond the temporal resolution of conventional tapping-mode MFM). The engineering solutions—longer on-chip waveguides and a low-profile microwave probe—overcame geometric constraints under the MFM head, enabling simultaneous electrical excitation/detection and magnetic imaging.
Conclusion
The authors present a method to probe the spatial profile of operational magnetic nanodevices by extending an MFM system with a microwave probe station, enabling simultaneous electrical/microwave access during MFM scanning. Using this setup, they imaged operational nanoconstriction-based SHNOs and correlated MFM contrast with simulations of the current-induced Oersted field. The approach is useful for extracting spatial information from functioning nano-oscillators. Future directions include developing a quantitative MFM (e.g., qMFM) with similar electrical access to allow quantitative comparisons across scans and devices, and employing faster tapping or alternative modalities to directly resolve magnetization oscillations. The method can be applied to new device geometries and materials to further elucidate their operational magnetic profiles.
Limitations
- The MFM measurements reported are not quantitative; thus, results from different scans cannot be directly compared in absolute terms.
- The tapping-mode temporal resolution (kHz) is insufficient to resolve GHz magnetization oscillations, preventing direct imaging of the oscillatory dynamics.
- Mechanical and spatial constraints under the MFM head required custom long waveguides and low-profile probing, which may influence device layout compared to conventional designs.
- The excerpted fabrication detail includes a partially truncated base pressure value; precise vacuum conditions may matter for reproducibility but are not fully specified here.
Related Publications
Explore these studies to deepen your understanding of the subject.