Levitated Milligram-scale Ferromagnetic Magnetometer at Room Temperature

AI summary

Overview

Research area: Quantum physics / precision metrology — magnetometry using optically detected, diamagnetically levitated mechanical oscillators (a levitated ferromagnet acting as a macrospin).

Technical level: Intermediate. The core idea is accessible, but the sensitivity budget (dissipation channels, thermal-noise limits, calibration) requires some comfort with power spectral densities and mechanical oscillators.

Scope (1 sentence): The paper reports a room-temperature, milligram-scale levitated NdFeB ferromagnet whose librational (rocking) motion is used to sense magnetic fields, reaching a measured sensitivity of 23 fT/√Hz near 153.001 Hz.

What This Paper Is About

Ultrasensitive magnetometers based on atomic vapors, SQUIDs and solid-state spins already reach subfemtotesla sensitivity, but better performance requires new physical platforms. A levitated ferromagnet is attractive because its many correlated electron spins are locked together, which in principle suppresses spin-projection noise and allows very high magnetic sensitivity — yet existing demonstrations (Meissner-levitated ferromagnets at 20 fT/√Hz) require cryogenics. The goal here is to build a comparably sensitive ferromagnetic magnetometer that works at room temperature, by optimizing the levitated object's geometry and systematically eliminating mechanical dissipation and noise.

Key Contributions

1. Demonstration of a room-temperature, milligram-scale (60.4 ± 0.1 milligram) diamagnetically levitated ferromagnet using a lifting magnet plus bismuth-nanoparticle plates, reaching a magnetic sensitivity of 23 fT/√Hz near the librational resonance at ω₀/2π = 153.001 Hz — stated by the authors to match the state of the art of cryogenic Meissner-levitated systems (20 fT/√Hz).
2. A dissipation-engineering strategy: an elongated 1 mm × 1 mm × 8 mm bar to minimize I/V², bismuth nanoparticles with native insulating oxide layers to confine eddy currents, a PEEK frame, and a two-layer MnZn ferrite shield — yielding a low-pressure damping of γ/2π = 0.49 mHz (Q = 3.1 × 10⁵) with a ringdown time of 324.3 s.
3. A systematic accounting of every loss channel (eddy currents in the ferromagnet, in external conductors, and in the Bi plates; magnetic hysteresis in the shield and in the ferromagnet), identifying magnetic hysteresis inside the levitated ferromagnet (γ_hyst^fm/2π ~ 0.1 mHz) as the dominant residual dissipation.
4. A dual-method, cross-validated calibration of the magnetic sensitivity: direct optical rotation-angle measurement and a deliberately off-resonant (360 Hz) magnetic excitation, which agree to within uncertainty — (1.9 ± 0.1) × 10⁻¹² N·m predicted versus (2.0 ± 0.1) × 10⁻¹² N·m measured.

Main Findings

• Measured sensitivity: 23 fT/√Hz near the 153.001 Hz librational mode, described as being at the "100-Hz level" and limited by vibration noise rather than thermal noise.
• Mechanical quality: At 8.6 × 10⁻² mbar the x-axis mode has a FWHM of γ/2π = 4.40 mHz (Q = 3.5 × 10⁴) and a ringdown time of 36.2 s; at 6.4 × 10⁻⁵ mbar the decay time rises to 324.3 s, giving γ/2π = 0.49 mHz and Q = 3.1 × 10⁵, verified to be pressure-independent.
• Two rotational modes observed: 28.126 Hz about the y-axis and 153.001 Hz about the x-axis; three translational modes below 5 Hz are not shown, and the z-axis mode has low sensitivity with the optical scheme used.
• Dissipation budget: Eddy currents contribute γ_eddy^fm/2π ≈ 0.01 mHz (ferromagnet), γ_eddy^out/2π ≈ 3 × 10⁻⁷ Hz (external conductors) and γ_eddy^Bi/2π ≈ 3 × 10⁻¹⁰ Hz (bismuth plates); hysteresis contributes γ_hyst^shield/2π ≈ 0.03 mHz (ferrite shield) and γ_hyst^fm/2π ~ 0.1 mHz (ferromagnet), the latter matching the measured total.
• Theoretical limits: The thermal-noise-limited sensitivity at low pressure is 2 fT/√Hz; the projected sensitivity after further dissipation reduction is 0.3 fT/√Hz.
• Calibration cross-check: An applied 360 Hz field (coil current amplitude 0.197 μA) produced B_cal = 0.29 ± 0.02 nT after a shield enhancement factor α_shield = 1.3, giving torques of (1.9 ± 0.1) × 10⁻¹² N·m theoretically and (2.0 ± 0.1) × 10⁻¹² N·m experimentally.
• Detector characterization: The optical readout is linear over ±0.2 mm (R² = 0.998, five repeated measurements) with response coefficient k_y = 937 ± 25 V/m; that range corresponds to ±2.6 × 10⁻⁴ rad, about four orders of magnitude above the room-temperature thermal fluctuation of 2.1 × 10⁻⁸ rad.
• Noise environment: Platform vibration noise is on the order of 10⁻⁶ g/√Hz near 100 Hz, and translational modes below 5 Hz are identified as the likely perturbation of the x-axis rotational mode.

Methodology in Plain English

The team suspends a small bar magnet in mid-air without mechanical contact. A permanent lifting magnet above cancels most of the bar's weight, while two plates of bismuth nanoparticles (held in cetyl alcohol) provide a diamagnetic repulsion that creates a stable potential well. Because the bar is a hard ferromagnet with its magnetization pointing up, any horizontal magnetic field twists it sideways; it rocks, or librates, about the field axis at a frequency set by its magnetization, the applied field, its volume and its moment of inertia.

A thin silicon strip glued under the bar acts as a mirror. A 532-nm laser reflects off it through a 0.38 m optical lever onto a four-quadrant photodetector, so the rocking angle becomes a voltage. The whole assembly sits inside a PEEK frame (an insulator, so no eddy currents), inside a two-layer ferrite magnetic shield, on a three-stage spring-mass isolation stack, with temperature held by a PID controller.

To find and reduce loss, the authors measure the noise power spectrum of the rocking angle over 7200 s (six 1200-s segments) and separately perform ringdown measurements at high and low vacuum. They then model each loss channel — eddy currents in the bar, in nearby conductors and in the bismuth, and magnetic hysteresis in the shield and in the bar itself — with finite-element simulations and compare to the measured damping. Finally, they validate the magnetic sensitivity in two independent ways: converting the measured angular noise into field noise using the known moment of inertia and response function, and applying a known off-resonance field to confirm the torque scale.

Why This Matters

This work shows that a room-temperature, tabletop-scale levitated magnet can rival the magnetic sensitivity of cryogenic levitated systems, removing the need for cryogenics while keeping a path to further improvement through materials and vibration isolation.

Real-world applications (as framed by the paper):

• Biomagnetic field detection, including fields at the 23 fT/√Hz noise floor reported here.
• Characterization of nanoscale magnetic phenomena.
• Searches for beyond-standard-model forces, including axions and dark photons and exotic spin-dependent interactions.
• Tests of couplings between single spins and rotations, and studies of rotational quantum superpositions and spin-mechanics coupling.

Industry relevance: the platform combines a permanent-magnet sensor head, optical readout and vacuum/electromagnetic shielding — a combination relevant to low-field sensing instrumentation, magnetic shielding and precision isolation engineering, and to anyone building compact magnetometers that must operate without cryogenics.

Future Directions

• Vibration-noise reduction. The authors identify vibration as the current limiting noise and propose upgrading the multi-stage isolation, raising the frequencies of the sub-5-Hz translational modes via bismuth-plate geometry engineering, and applying feedback cooling to those modes, since passive isolation is ineffective at low frequency.
• Lower-loss materials and geometry. Hysteresis loss scaling (∝ μ″(a² + b²)ω₀³/M²) suggests reducing the imaginary permeability, lowering ω₀, minimizing transverse dimensions, and increasing magnetization, aiming at ~10 μHz dissipation (Q ~ 10⁷ at 100 Hz).
• Pursuing 0.3 fT/√Hz and below at room temperature, with the platform's reported cryogenic compatibility offered as a further route to higher sensitivity.
• Fundamental-physics campaigns. Deploying the sensor in source–sensor experiments for exotic spin-dependent interactions, dark-matter searches, and rotational quantum-state studies.

Target Audience

Researchers in precision metrology, optomechanics and levitated sensing; quantum-physics experimentalists interested in spin-mechanics coupling and rotational quantum states; instrument developers working on biomagnetic or low-field magnetic sensing; and graduate students who want a clear worked example of how a mechanical oscillator's dissipation budget is measured, decomposed and pushed toward the thermal limit.

Authors’ abstract

Levitated mechanical oscillators are emerging ultrasensitive sensors with tremendous potential in both applied and fundamental physics. Levitated ferromagnets, with internal spin noises rapidly averaged, promise ultrahigh magnetic sensitivity. Here, we demonstrate a milligram-scale diamagnetically levitated ferromagnet system operating at room temperature. Through optimized geometry and multi-channel dissipation control, we achieve a magnetic sensitivity of 23~fT$/\sqrt{\text{Hz}}$ at frequency of 100-Hz level. We anticipate that a ferromagnetic magnetometer with subfemtotesla sensitivity is within reach, after modest technical improvements. This platform establishes a high-performance magnetometer for biomagnetic field detection and beyond-standard-model force searches.