Skip to content

Fiber-integrated N - V diamond magnetometer compatible with commercial endoscopic systems

FieldValue
JournalPhysical Review Applied
Published2026-03-10
AuthorsSatbir Singh, Hyunjong Lee, Nhu Anh Nguyen, Seonghyeon Kang, Jeong Hyun Shim
DOI10.1103/qhwp-kf51

A 10 mm diameter endoscopic head achieved a magnetic field noise floor of approximately 3 nT/√Hz using continuous wave magnetometry. The same assembly reached a 0.85 nT/√Hz noise floor when an external magnet provided a more uniform bias field. These results show that the bulk N-V configuration supports a compact optical probe, although magnet placement limited the measured noise floor.

The sensing element was a (100)-cut single-crystal diamond, DNV-B14, with approximately 4.5 ppm N-V centers. It was attached to an FP1500ERT multimode fiber with a 1.5 mm core diameter and numerical aperture of 0.50. The same fiber carried excitation and collected photoluminescence, so collection efficiency set a direct limit on sensitivity.

The authors shaped both parts to reduce optical loss. The fiber core and cladding had refractive indices of 1.458 and 1.365. Their critical angles were 69.5° at the core/cladding interface and 43.3° at the core/air interface. The paper modeled diamond emission at 640 nm, with refractive index about 2.42, over angles from 0 to 90° toward the fiber core. This angular mismatch caused total internal reflection losses in parts of the fiber.

The fiber facet was ground into a curved geometry near the tip. The surface angle changed gradually from 0° to 43.3°, while removal of the thin cladding enabled a lower critical-angle condition at the core/air interface. The apex diameter decreased from 1.5 mm to 1.0 mm. The diamond slab began as 1.5 × 1.5 × 0.5 mm³. It was polished into a truncated cone with a 1.0 mm base diameter, 0.5 mm height, 0.5 mm top diameter, and about 25° semi-vertical angle. The 25° angle was close to the 24.6° diamond/air critical angle.

The shaped geometry was intended to reduce photoluminescence trapped inside the diamond and direct it into the fiber. In the optical test, the fiber end collected a signal about 12 times stronger than collection through the 0.6 NA air objective. COMSOL ray tracing with 2000 isotropic rays estimated about 23% collection through the fiber. The paper cited about 2% collection for a diamond slab with a 0.6 NA air objective. It reported no more than 10% with a 1.45 NA oil objective. The 23% estimate came from simulation, whereas the 12-fold result was measured.

The authors aligned the bias field with one of the four N-V axes to isolate one orientation. They used a 3D-printed head with a channel for the instrument tube and a slot for a ring-shaped magnet. The magnet had a 3 mm diameter, 1 mm width, and N50 neodymium grade. The slot angled the magnet 54.74° to the vertical axis, matching the selected N-V orientation.

The commercial clamp tube had approximately 4 mm external and 3 mm internal diameters, matching a conventional endoscopic channel. The fiber was secured in the tube with heat-shrink tubing. This prevented unintended movement while allowing manual rotation and axial translation for alignment. The 10 mm head limited magnet separation, so the permanent magnet sat close to the diamond.

A nonmagnetic coaxial cable supplied microwave excitation for the $m_s = 0$ and $m_s = \pm 1$ transitions. Its inner conductor was wrapped around the tube near the diamond, and the tube protected the wire from contact. The dedicated slot fixed the magnet angle, while manual fiber alignment and heat-shrink retention provided mechanical stability.

The optical setup used a 532 nm continuous wave laser with 300 mW power at the portable 30 cm × 30 cm setup. A dichroic mirror separated excitation from returning photoluminescence, and a 650 nm long-pass filter selected the N-V emission. Two photodetectors operated at equal gain for balanced detection. Microwave frequency modulation and a lock-in amplifier measured the ODMR slope.

For the integrated-magnet configuration, the modulation frequency was 5 kHz and the depth was 8.5 MHz. The resulting spectrum showed a clear ODMR signal for a well-isolated [111] axis. A linear region was available for the $m_s = 0$ to $m_s = -1$ resonance. The lock-in time constant was 1 ms, corresponding to 125 Hz bandwidth with 12 dB/octave settings. The in-phase output was digitized for analysis.

A Helmholtz coil applied a known alternating field during sensitivity testing. The coil produced 1.73 μT RMS at 20 Hz along [100] at the diamond. Its component projected onto [111] was 1 μT RMS. The coil was calibrated against the field inferred from ODMR resonance shifts. A 1 kS/s trace was recorded for 1 s within the linear slope region. FFT analysis then produced the RMS amplitude spectral density and the magnetic noise floor.

The endoscopic magnetometer configurations exhibited distinct noise floors under identical test conditions, emphasizing the sensitivity limitations of the integrated design and the impact of magnetic field uniformity.

ConfigurationBias field sourceApplied test fieldMagnetic noise floor
A, integrated headN50 neodymium magnet1 μT RMS projected onto [111] at 20 Hzapproximately 3 nT/√Hz
B, external magnetLarger magnet at sufficient distance1 μT RMS projected onto [111] at 20 Hzapproximately 0.85 nT/√Hz

These measurements confirm that separating the bias magnet from the sensing element substantially reduces the noise floor, directly linking magnetic field uniformity to improved sensitivity.

In a confocal arrangement, a DNV-B14 slab with the same specification reached approximately 1 nT/√Hz. The endoscope therefore had a higher noise floor. The authors attributed the difference to ODMR linewidth broadening caused by magnetic field nonuniformity over the illuminated area.

Earlier fiber-integrated magnetometers reported sub-nT sensitivity [22,23], and a portable integrated system also reached sub-nT performance [27]. An ensemble N-V magnetometer has reported 0.5 pT/√Hz [18]. The present approximately 3 nT/√Hz result is lower than those prior values. The paper frames the compact, mechanically stable configuration as a feasibility study for difficult-to-access regions.

Configuration B replaced the close integrated magnet with a larger diameter permanent magnet at a sufficient distance. The authors assumed that this arrangement provided a uniform field at the diamond. The resulting ODMR spectrum showed clear hyperfine splitting, which was absent in configuration A. The optimized modulation depth was about 0.6 MHz.

Under the same 20 Hz test field, configuration B produced a measured noise floor of about 0.85 nT/√Hz. That value came from the external-bias configuration rather than the compact integrated head. The authors suggest optimizing the permanent magnet configuration to reduce linewidth broadening in the integrated design.

They also expect an approximately fourfold further improvement through simultaneous hyperfine-transition driving and double-resonance excitation. This projected gain was not demonstrated in the reported measurements. Configuration B suits test environments that allow an external bias field. Configuration A remains useful where endoscopic use rules out an external magnet.

For diamond selection, the paper used a bulk single-crystal N-V element with a defined cut, defect concentration, and optical geometry. The truncated cone and curved fiber increased collected light, while magnetic nonuniformity from the close magnet limited sensitivity. Sensor engineers must balance photoluminescence coupling, field uniformity, and probe diameter when translating this design to other instruments.

The 4 mm instrument tube fit inside the 10 mm head, leaving room for monitoring or cleaning channels proposed by the authors. The 30 cm × 30 cm portable optical setup remained separate from the 10 mm head. The compact configuration trades measured sensitivity for operation without an external magnet. All reported sensitivity values came from continuous wave measurements. The authors leave permanent magnet optimization to future work. The demonstrated platform supports magnetic-field measurements in difficult-to-access regions.

The paper needs a shaped and polished (100)-cut single-crystal diamond with N-V centers for a fiber-coupled endoscopic magnetometer. 6CCVD’s diamond materials catalog lists MPCVD-grown single-crystal diamond with thickness from 0.1 to 500 micrometers and polishing to Ra below 1 nm. The 0.5 mm starting slab fits that range, and the linked engineering support page can discuss custom fabrication of the diamond geometry.

View Original Abstract

The nitrogen-vacancy ( <a:math xmlns:a=“http://www.w3.org/1998/Math/MathML” display=“inline”> <a:mrow> <a:mrow> <a:mi mathvariant=“normal”>N</a:mi> </a:mrow> </a:mrow> </a:math> -) center in diamond provides a robust, solid-state platform for magnetic field measurements at room temperature. To harness its potential in inspecting inaccessible regions, here we present a compact endoscopic configuration of a <d:math xmlns:d=“http://www.w3.org/1998/Math/MathML” display=“inline”> <d:mrow> <d:mrow> <d:mi mathvariant=“normal”>N</d:mi> </d:mrow> </d:mrow> </d:math> - diamond-based magnetometer. The endoscopic magnetometer was developed by integration of a large-core optical fiber with a bulk <g:math xmlns:g=“http://www.w3.org/1998/Math/MathML” display=“inline”> <g:mrow> <g:mrow> <g:mi mathvariant=“normal”>N</g:mi> </g:mrow> </g:mrow> </g:math> - diamond for laser excitation and photoluminescence (PL) collection. The diamond and fiber were specially shaped to enhance PL collection through the fiber. Additionally, a three-dimensionally-printed endoscope head was used to facilitate alignment of the bias magnetic field along the <j:math xmlns:j=“http://www.w3.org/1998/Math/MathML” display=“inline”> <j:mrow> <j:mrow> <j:mi mathvariant=“normal”>N</j:mi> </j:mrow> </j:mrow> </j:math> - axis. A magnetic field sensitivity of approximately 3 nT/√Hz at 20 Hz was achieved by continuous-wave magnetometry measurements. The endoscope diameter was restricted to 10 mm to match the dimensions of most commercial endoscopes. The magnetic field nonuniformity caused by the small separation between the diamond and the magnet in the endoscope head limited the overall sensitivity. It could be further improved to 0.85 nT/√Hz at 20 Hz by use of a magnet placed at a sufficient distance outside the endoscope head. Our endoscopic design is mechanically stable and provides additional opportunities for integrating other functionalities into the probe head as needed.