Nanoscale SQUIDs: Fabrication & Applications

• Nanoscale SQUIDs are ultra-sensitive magnetometers that employ quantum interference in superconducting loops with weak links to detect minute magnetic fields.
• They integrate advanced fabrication methods such as FIB milling, corner lithography, and angled evaporation to achieve sub-100 nm device dimensions and on-tip circuit integration.
• These devices offer exceptional performance with noise floors below 1 μΦ0/√Hz and are used for high-resolution imaging of complex magnetic nanostructures in quantum materials.

Nanoscale superconducting quantum interference devices (SQUIDs) are highly miniaturized, ultra-sensitive magnetometers capable of detecting minute magnetic fields at nanometer spatial resolution. These devices exploit quantum interference in a superconducting loop interrupted by weak links—typically Dayem-bridge Josephson junctions (JJs)—to transduce external magnetic flux into an electrically measurable quantity. The integration of such SQUIDs onto scanning probe platforms, including sharp tips and compliant cantilevers, has enabled direct, noninvasive imaging of complex magnetic nanostructures, mesoscopic phenomena, and correlated electron states across a range of material systems, all while preserving high spatial and magnetic field sensitivity. The field has advanced rapidly in recent years by leveraging lithographic miniaturization, focused-ion-beam (FIB) patterning, advanced wafer-scale methodologies, and self-alignment strategies, culminating in device diameters below 100 nm, flux noise floors below $1~\mu\Phi_0/\sqrt{\mathrm{Hz}}$, and compatibility with extreme sample environments including high fields and low temperatures (Weber et al., 3 Aug 2025, Roskamp et al., 16 Jan 2026, Wyss et al., 2021, Finkler et al., 2010).

1. Fabrication Workflows and Device Architectures

Nanoscale SQUIDs are fabricated using a variety of self-aligned techniques optimized for robust tip or cantilever integration:

• Wafer-scale planar cantilever approaches: Utilized by Nb-based SQUID-on-lever devices, this workflow involves starting with a silicon-on-insulator (SOI) wafer, growing a thin SiO$_2$ protection layer, and sputtering films of Nb (typically 50–60 nm) with capping layers (Al$_2$O$_3$, Pd, Au). Standard optical lithography and reactive-ion etching (RIE) define cantilever outlines and device geometry, followed by deep RIE for chip release and hydrofluoric acid (HF) for final cantilever freeing. At the cantilever apex, FIB milling—using Ne, He, or Ga ions—precisely sculpts SQUID loops and Dayem-bridge JJs, with loop diameters as small as 10–15 nm (minimum) and Dayem-bridge widths down to 10 nm using He-FIB (Weber et al., 3 Aug 2025).
• Corner lithography wireframe platforms: Molding and corner lithography on Si(100) wafers generate inverted pyramids which, after conformal LPCVD nitride deposition, selective isotropic etching, and shadow-mask sputtering, yield wireframe apex tips supporting self-aligned Nb nanowires. FIB nanopatterning at the tip apex forms Dayem-bridge weak links. Device diameters are tunable from several microns down to 100 nm, with batch scalability and precise apex alignment (Roskamp et al., 16 Jan 2026).
• SQUID-on-tip by angled metal evaporation: Pulled quartz tubes with sharpened apices (down to 100 nm diameter) are coated using three-angle aluminum evaporation, naturally forming two superconducting leads and an apex ring with two thin arcs serving as Dayem-bridge JJs. This method achieves device sizes down to $d = 100$ nm, requires no lithography, and delivers high on-tip field compatibility (Finkler et al., 2010).
• FIB-defined SQUIDs on commercial AFM cantilevers: Sputter-deposited multilayer films (e.g., Ti/Nb/Pt) on silicon AFM cantilevers are FIB-milled into triangular plateaus, upon which are patterned SQUID loops (e.g., effective diameter $D_{\text{eff}} = 365$ nm), Dayem bridges, and integrated shunt resistors for flux and thermal sensing (Wyss et al., 2021).

These approaches share a focus on minimizing process misalignment, maximizing reproducibility (self-alignment to apex), and allowing for device miniaturization far beyond conventional planar SQUIDs.

2. Device Principles, On-Tip Circuits, and Readout

Nanoscale SQUIDs operate fundamentally as flux-to-voltage or flux-to-current transducers based on the quantum interference of the superconducting order parameter in a loop interrupted by one or more weak links. Essential parameters include the loop area $A_{\mathrm{eff}}$, junction critical current $I_c$, and loop inductance $L$, the latter dominated by kinetic inductance in nanowires and thin films.

Advanced on-tip integration encompasses:

• 2JJ and 3JJ geometries: Two Dayem-bridge JJs with an adjacent modulation line for flux-biasing constitute the 2JJ-SoL design. The addition of a third junction (3JJ-SoL) provides phase-bias capability, allowing shifted $I_c(B)$ interference patterns and reproducible working-point control without large stray fields. Key parameters include mutual inductance $M = \Phi_0/\Delta I_{\mathrm{mod}}$, screening parameter $\beta_L = 2LI_0/\Phi_0$, and phase tuning inductance $L_{\mathrm{ctrl}}\ll M$ for efficient phase control ($I_{\mathrm{ctrl}}\sim$ tens of $\mu$A) (Weber et al., 3 Aug 2025).
• On-chip resistor integration: FIB-induced Pt deposition creates shunt resistors (e.g., 4 $\Omega$, 20–30 $\mu$m$^2$ area) at the tip, enabling thermal imaging functionality via temperature-dependent current response (Wyss et al., 2021).
• Wireframe tip integration: Corner lithography templates support additional local field coils and flux modulation leads, increasing on-tip circuit complexity while retaining sub-100-nm feature control (Roskamp et al., 16 Jan 2026).

Electrical characterization employs both voltage- and current-bias readouts, with cold SQUID series-array amplifiers and flux-locked loop circuits common for maximizing bandwidth and linear response.

3. Performance Metrics and Key Figures of Merit

Recent advances in device miniaturization, JJ engineering, and readout optimization have established a new performance frontier for nanoscale SQUIDs:

Metric	Nb SoL (3JJ, (Weber et al., 3 Aug 2025))	Wireframe (Roskamp et al., 16 Jan 2026)	Nb Cantilever (Wyss et al., 2021)	Al SOT (Finkler et al., 2010)
Loop diameter	15–87 nm	114–833 nm	365 nm	100–400 nm
Dayem-bridge width ($w$)	10–40 nm	80–120 nm	50 nm (thk 50 nm)	30 nm (thk 17 nm)
Loop inductance ($L$)	1–5 pH	20 pH	—	550 pH (kinetic dom.)
Screening parameter ($\beta_L$)	0.10	0.53	0.1–0.2	0.85
Flux noise $S_\Phi^{1/2}$	0.3 $\mu\Phi_0$/√Hz	3.8 $\mu\Phi_0$/√Hz	0.48 $\mu\Phi_0$/√Hz	1.8 $\mu\Phi_0$/√Hz
Field sensitivity $S_B^{1/2}$	120 nT/√Hz	—	9.5 nT/√Hz	0.11 $\mu$T/√Hz
Spatial resolution (PSF FWHM)	87 nm	100–800 nm	100–200 nm (mag/thermal)	100 nm (practical)
Max. field operation	0.5 T	$\geq$1 T	1.0 T	0.6 T
Operating temperature	4.2 K	3–7 K	4.2 K	0.3 K

Performance context: Loop miniaturization improves spatial resolution and dipole sensitivity for nanoscale magnetic imaging, while sufficient mutual inductance ($M$) and $\beta_L$ ensure strong transfer functions without excessive noise or nonlinearity. Field operation up to 1 Tesla, white-noise floors below 1 $\mu\Phi_0/\sqrt{\mathrm{Hz}}$, and stability at cryogenic temperatures are common features.

4. Nanoscale Magnetic Imaging Applications

Nanoscale SQUIDs, particularly when integrated with scanning probe platforms, enable a broad range of highly resolved, quantitative magnetic imaging modalities:

• Topographic-integrated magnetic mapping: Leveraging compliant cantilevers and simultaneous atomic force microscopy feedback, tip-sample distance can be maintained with sub-nanometer vertical sensitivity and lateral resolution down to 50 nm (Wyss et al., 2021).
• Emergent spin textures: Imaging experiments, such as direct visualization of skyrmions and helical magnetic phases in Cu$_2$OSeO$_3$, demonstrate spatial resolution sufficient to resolve 95 nm skyrmion separations and magnetic modulations with a period as small as 65 nm. Simulation-aided deconvolution provides empirical point-spread functions (PSF) with FWHM $\sim$87 nm (Weber et al., 3 Aug 2025).
• Quantitative magnetometry and spectroscopy: The devices' calibrated flux response and miniaturized pickup loops allow for quantitative mapping of stray fields, local critical currents, and magnetization modulations, with relevance to studies of 2D van der Waals magnetism, correlated electron phenomena, superconducting vortices, and topological insulator physics (Wyss et al., 2021, Finkler et al., 2010).
• Thermal imaging: On-tip integrated shunt resistors enable dual magnetic and thermal sensing, with demonstrated thermal sensitivities of 620 nK/√Hz, facilitating visualization of dissipative processes at the nanoscale (Wyss et al., 2021).
• Single-spin sensitivity: For devices with minimized loop area and weak link width ($w$), projected spin sensitivity reaches $<20~\mu_B$/√Hz for on-axis dipoles adjacent to the loop (ring diameters down to 130 nm) (Finkler et al., 2010).

5. Scalability, Wafer-Scale Manufacture, and Integration

Recent developments emphasize the transition from single-device demonstrations toward scalable, wafer-scale manufacture:

• Batch fabrication: Corner lithography and shadow-effect deposition enable the production of hundreds of wireframe SQUID-on-cantilevers per 4-inch wafer, with all critical patterning steps completed in parallel. Apex geometries, loop size, and wire width are tunable via photolithographic mask openings, conformal film thickness, and FIB parameters (Roskamp et al., 16 Jan 2026).
• Self-alignment and drift-corrected FIB: Lithographic referencing of tip apices and active beam-position correction during FIB milling permit sub-10-nm alignment accuracy, ensuring consistent spatial registration of the SQUID with the probe’s scanning axis (Weber et al., 3 Aug 2025).
• Monolithic on-tip circuit integration: Elaborate on-chip designs are feasible, embedding modulation lines, phase tuning leads, bias/field coils, and shunts in the probe head itself. Such integration supports local flux control, phase biasing, and susceptibility measurements (Roskamp et al., 16 Jan 2026, Weber et al., 3 Aug 2025).
• Hybrid AFM–SQUID platforms: By integrating SQUIDs with compliant AFM levers, the combined platform supports simultaneous, high-resolution topographic and magnetic imaging in a single scan, overcoming limitations of tuning fork-based approaches (Wyss et al., 2021).

6. Comparative Assessment and Future Prospects

Approach	Minimum Loop Diameter	Pros	Cons / Limitations
SQUID-on-tip (Al, quartz) (Finkler et al., 2010)	100 nm	No lithography, high $\mu_B$-sensitivity	Limited circuit complexity, feedback via tuning fork only
FIB SoL on planar lever (Weber et al., 3 Aug 2025)	10–15 nm	AFM integration, 3JJ design, high spatial resolution	Fabrication complexity, limited by FIB scope
Wireframe/corner-lith. (Roskamp et al., 16 Jan 2026)	100 nm	Batch scalability, circuit integration	KOH/Si etch intricacy, apex control
FIB on AFM cantilever (Wyss et al., 2021)	365 nm	Robust topography/magnetic/thermal	Larger minimum loop diameter

Potential improvements include further loop miniaturization via He$^+$ FIB or tip-specific angled-evaporation, lower-noise and higher-$T_c$ superconductors (e.g., Pb, YBCO), and expanded on-chip integration (local compensation, feedback). A plausible implication is the extension of quantitative SQUID imaging to quantum materials where invasive probes or coarse spatial resolution have previously been prohibitive.

7. Conclusion

Nanoscale SQUIDs constitute a unique measurement paradigm at the interface of quantum physics, nanofabrication, and scanning probe instrumentation. The current state of the art delivers sub-100-nm spatial resolution, sub-$\mu\Phi_0$/√Hz noise floors, and robust operation in applied fields up to 1 T. Self-aligned fabrication approaches, wafer-scale scalability, and on-chip circuit complexity collectively enable advanced functional imaging of emergent magnetic and thermal phenomena. The field is converging toward universal, robust, and scalable platforms capable of addressing grand challenges in quantum magnetism, correlated electron systems, and nanoscale materials science (Weber et al., 3 Aug 2025, Roskamp et al., 16 Jan 2026, Wyss et al., 2021, Finkler et al., 2010).