Skip to content

Magnetic field imaging with nanodiamond-based quantum sensor probe

FieldValue
JournalAIP Advances
Published2026-09-01
AuthorsTanmay Talukder, Sheikh Shahparan Mahtab, Peker Milas, Tomas Sujeta, Rajib Saha
DOI10.1063/5.0324550

A glass micropipette coated with commercially available 140 nm nanodiamonds mapped a 0.69 to 1.44 mT magnetic field across a 1.95 × 3.12 mm² region. The region contained fragmented neodymium permanent magnets. The probe used 40 continuous-wave ODMR scans and gave an estimated sensitivity of about 120 μT/√Hz at a signal-to-noise ratio of about 3. This is a low-cost imaging platform with sub-millimeter to millimeter mapping scale. It is not a demonstrated nanoscale or full-vector magnetometer.

The sensing head was made from a 6 in. glass capillary with a 1 mm outer diameter and a 0.42 mm wall. A Sutter Instruments P-97 puller formed the taper, and the tip was broken to enlarge its opening to about 5 μm. Carboxyl-coated nanodiamonds with an average diameter of 140 nm were then immobilized on the outer taper. The authors achieved deposition by immersing the micropipette in the nanodiamond solution and drying it in air ten times.

The test sample contained fragmented neodymium permanent magnets fixed on a bare silicon wafer and covered with cured polydimethylsiloxane, or PDMS. Its patterned area was 1.95 × 3.12 mm². The probe remained 1 mm above the sample while an XYZ stage with micrometer controllers moved the specimen.

Nanodiamonds on the micropipette produced more photoluminescence than nanodiamonds on flat glass or silicon. The paper attributes this to possible optical field redistribution and resonant effects from the curved tip. The mechanism was not established through quantitative electromagnetic or optical analysis.

The probe needed a 2.87 GHz field to drive transitions between the NV center spin sublevels. The authors used Ansys HFSS simulations to design a copper wire antenna and maximize the simulated magnetic field near the nanodiamond tip. The initial straight-wire length was set to 10.45 cm, based on the quarter wavelength at the NV zero-field splitting.

An earlier tungsten-wire design did not produce ODMR, so the authors separated the glass probe from the copper antenna. The fabricated antenna used 0.3 mm copper wire in a spring-shaped body with a U-shaped end and a female SMA connector.

Its measured return loss was -20.7 dB at 2.87 GHz. The authors reported a reflected-power fraction of about 0.85% and an impedance-mismatch delivery efficiency of about 99.15%. For a simulated input of 1 W, the upper U-shaped region produced an estimated magnetic-field conversion efficiency of 0.24 to 0.35 mT/√W.

The antenna concentrated the microwave field near the sensing region, but it was not optimized for uniformity over an extended area. The authors state that field inhomogeneity may have broadened ODMR lines and reduced contrast in ensemble measurements.

A custom confocal setup operated at room temperature. A 532 nm DPSS laser excited the defects through a 50× long-working-distance objective with NA 0.42. The measured focal spot was about 5 μm in diameter, compared with a theoretical diffraction limit of about 1.2 μm. A deliberate underfill increased the effective focal volume and helped maintain alignment and light throughput.

Microwaves passed through the antenna from an HP Agilent 8648C generator and a MiniCircuits ZVA-02443HP+ amplifier. An ASEQ Instruments HR1 CCD spectrometer recorded photoluminescence, and a Newport 1830C power meter checked laser stability. Scans began only after the laser power fluctuated by no more than 1%.

An expectation-maximization machine-learning method removed background noise from the spectra. Each continuous-wave spectrum covered 100 microwave frequencies at 350 ms per point, for a minimum acquisition time of 35 s. The magnetic-field update rate was about 0.029 Hz. The theoretical Nyquist limit was 0.014 Hz, while sequential acquisition made the practical DC bandwidth about 0.01 Hz or lower.

The integrated photoluminescence was about 1.9 × 10⁷ counts per spectrum, with an estimated detected photon count rate of 5.4 × 10⁷ counts s⁻¹. Fitting two Lorentzian dips over a linear baseline gave linewidths of about 20 MHz. The ODMR signal-to-noise ratio was about 3, so the authors treated their sensitivity as an estimate under continuous-wave conditions rather than an optimized, shot-noise-limited value.

The sample was divided into a 5 × 8 grid with 40 evenly spaced positions. Each cell covered about 400 × 400 μm². At every position, the authors corrected the ODMR contrast with a polynomial fit to data below 2.84 GHz and above 2.89 GHz. A cubic spline then located the two minima around the NV resonance.

The minimum separation was converted to magnetic field with Δf = γeB. A calibrated three-axis Helmholtz coil verified the field values. The reconstructed map gave 0.69 to 1.44 mT, with higher values aligned with regions containing magnetic material. The raw ODMR contrast averaged 2.3% across the scans and ranged from 1.75% to 3.5%.

Because the NV centers had random orientations and no vector reconstruction algorithm was used, the probe measured a scalar magnetic-field projection. It did not reconstruct all three field components.

The optical excitation volume was micrometer scale, but the magnetic map remained sub-millimeter to millimeter scale. The main constraints were the 1 mm probe-to-sample separation, the 400 μm sampling interval, and the 5 μm tip diameter. The tip collected enough light from an ensemble of nanodiamonds, but its larger collection area reduced lateral selectivity.

The 50× objective also had a relatively low NA of 0.42. Low collection efficiency and optical losses in the detection path prevented measurements from single nanodiamonds or smaller ensembles. The authors proposed smaller tips of 500 nm or less, a lower-working-distance objective with higher NA, and higher-efficiency detectors such as avalanche photodiodes.

Full vector imaging would require controlled NV orientation and multi-axis ODMR analysis. Random orientation, strain inhomogeneity, and surface defects also produced broader lines, lower contrast, and lower sensitivity than optimized single-NV or bulk-diamond systems.

The reported sensitivity of about 120 μT/√Hz provided less magnetic-field sensitivity than the paper’s cited nanodiamond result of 570 pT/√Hz, which used lock-in detection. It also trailed the femtotesla sensitivity reported for bulk diamonds with pulsed ODMR and lock-in methods. Different readout methods and hardware prevent a direct defect-level comparison.

A cavity-based microwave system previously described by Le Floch and colleagues offers better field uniformity and power efficiency than the simple antenna used here. The present design instead favored low-cost fabrication and proof-of-principle spatial mapping. Its nonuniform microwave field was accepted as a trade-off.

The authors suggest adding pulsed ODMR and lock-in amplification to the micropipette platform. They also plan larger datasets for machine-learning analysis. Earlier collection from individual 35 nm nanodiamonds supports the stated route toward a single-nanodiamond probe, but that configuration was not tested in this work.

The paper shows that a tapered glass micropipette can carry a functional NV-nanodiamond sensing head without the complex cantilever fabrication used for bulk-diamond scanning probes. The result is magnetic mapping over a millimeter-scale region using commercial nanodiamonds and a simple copper antenna.

For device development, the measured limits point to specific design targets. Tip diameter and standoff set the spatial resolution. Photon collection sets the usable ensemble size and contrast. Antenna geometry sets spin-control uniformity. NV orientation determines whether the probe returns a scalar projection or a full vector field.

The result supports a proof-of-concept for low-cost ambient magnetic-field imaging. It does not match the sensitivity or spatial resolution of optimized bulk NV magnetometers.

The quantum sensor requires NV-center-containing, carboxyl-coated nanodiamond with an average diameter of 140 nm, immobilized on a tapered glass micropipette, and a copper microwave antenna. 6CCVD’s diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond, with Cu among its metallization options. The catalog does not specify this 140 nm nanodiamond format or its NV activation, so the exact material and custom integration requirements should be confirmed through engineering support.

View Original Abstract

Imaging and sensing of local magnetic fields are essential for advanced material characterization, providing insights into magnetic and electronic behaviors across multiple spatial scales. The electronic spin state of the negatively charged nitrogen-vacancy (NV) center in diamond is coupled to its luminescence, enabling local magnetic field detection via the Zeeman effect. Due to its atomic-scale dimensions and high sensitivity, the NV center serves as an ideal probe for sensitive magnetic field detection and imaging across multiple spatial scales. Although high-resolution magnetic field imaging using bulk diamond-based cantilever probes has been demonstrated, such approaches are limited by complex and costly fabrication processes. Here, we present a cost-effective method for magnetic field detection and imaging using a sensor probe fabricated by depositing 140 nm nanodiamond crystals on the tip of a glass micropipette. The increased photoluminescence intensity improved the optically detected magnetic resonance (ODMR) visibility and enabled stable ODMR acquisition under the present optical excitation conditions. A copper wire configured as a full wave radio frequency antenna efficiently delivered microwaves to the NV defects, enabling quantum-based sensing using this probe. The probe was used to image the magnetic field distribution over a 1.95 × 3.12 mm2 substrate containing permanent magnet fragments, sampled across a 5 × 8 grid. Spatial variations in the magnetic field were clearly resolved in the reconstructed image. This work demonstrates a proof-of-concept low-cost nanodiamond-based magnetic imaging platform for magnetic field detection and millimeter scale magnetic field mapping using nanodiamond-based quantum sensors.