Skip to content

Sol-gel processing meets quantum sensing - integrating NV--containing nanodiamonds into a fiber-based magnetic field sensor

FieldValue
JournalJournal of Sol-Gel Science and Technology
Published2026-03-24
AuthorsM. Weigel, A. Albrecht, J. Beyer, K. Kraushaar, C. Laube
DOI10.1007/s10971-026-07136-w
Citations1

A glass fiber tip coated with an NV-containing nanodiamond sol-gel layer produced a 15% decrease in photoluminescence at 50 mT. The response saturated after 50 mT within a test range of 0 to 90 mT. Specific sensitivity reached up to 0.4% PL intensity suppression per mT between 10 and 20 mT.

MeasureResultCondition
Applied-field test0 to 90 mTIron-core electromagnet
PL intensity change15% decreaseAt 50 mT
Specific sensitivityUp to 0.4% per mTBetween 10 and 20 mT

The measured 15% contrast exceeded the 12.5% maximum reported by Staacke et al. for pressed nanodiamond pellets over the same magnetic field range. The authors suggest that the higher contrast comes from a dense, optically accessible nanodiamond ensemble at the fiber tip.

More sensitive NV systems exist, but the authors note that they require oriented single-crystal diamond, single NV centers, or added microwave or background magnetic fields. The reported sensor avoids those requirements. It uses a microwave-free, all-optical readout and places the sensing material several centimeters from the excitation and detection equipment.

Sol-gel formulation and room-temperature curing

Section titled “Sol-gel formulation and room-temperature curing”

The D50 formulation replaces 50 mol% of methyltriethoxysilane, or MTES, with dimethyldiethoxysilane, or DMDES. This substitution lowered the degree of crosslinking and was intended to improve coating flexibility for optical fibers. Tetraethyl orthosilicate was deliberately avoided because its four alkoxide groups generally produce more brittle coatings.

PreparationMass ratioProcessing
SolMTES:DMDES:GPTES:0.1 M HCl = 9.5:8:1:2Stirred at room temperature for approximately 4 h
HardenerEtOH:GPTES:TBOT = 2:2:1Stirred for 1 h at room temperature
Coating mixtureSol:hardener = 1:1Nanodiamonds dispersed before drop coating

GPTES denotes (3-glycidyloxypropyl)triethoxysilane. TBOT denotes titanium(IV) n-butoxide. Ethanol could be added to the final mixture when lower viscosity improved handling.

Acid catalysis hydrolyzed the silane precursors, after which condensation formed Si-O-Si linkages. TBOT and GPTES acted as additional crosslinkers and allowed curing without thermal post-treatment. The coated fibers then dried at room temperature in ambient air for 48 h.

D50 was selected from earlier optimization work that included abrasion, cross-cut adhesion, peel adhesion, and cylindrical mandrel bend tests. It gave the best overall balance of coating flexibility and adhesion among the tested formulations.

Nanodiamond loading and coating construction

Section titled “Nanodiamond loading and coating construction”

The nanodiamonds came from Pureon and consisted of monocrystalline particles between 0.75 and 1.25 μm. As-received particles were used for coating optimization and optical characterization.

Loading was varied between 4 and 74 wt%. At 4 wt%, isolated nanodiamond agglomerates covered only small areas and could not ensure efficient excitation-light absorption. At 25 wt%, the particles nearly covered the surface. The upper loadings produced a tightly covered surface.

No visible cracking or delamination appeared in any tested layer. The maximum feasible loading was 74 wt%, because higher concentrations increased the sol viscosity enough to prevent uniform, high-quality coatings. The authors selected 25 wt% because it provided sufficient NV PL while limiting material consumption.

Before magnetic-field testing, the nanodiamonds were irradiated to create additional NV centers. The paper does not report the irradiation dose or fluence. A single laboratory pipette droplet of the coating mixture was placed directly on a stripped multimode glass fiber. The fiber had a 2 m length and a 105 μm core diameter.

Optical microscopy estimated the coating thickness at just below 10 μm. Local elevations reached up to 20 μm where the manually applied droplet formed thicker regions.

The D50 coating without nanodiamonds had a UV-Vis absorption edge below 350 nm. The authors attributed its transparency to nanoscale homogeneity, low surface roughness, and a lack of light-scattering domains.

The NV spectra retained the same features across all tested loadings. The zero-phonon line appeared at 638 nm, while phonon sidebands formed a broad emission band around 690 nm. Integrated PL intensity over 600 to 800 nm increased linearly with nanodiamond concentration and did not saturate over the investigated range.

The authors linked this linear response to fine particle dispersion and the absence of large agglomerates. These conditions limited scattering and excessive absorption of both the excitation light and emitted PL.

Thermal analysis found decomposition beginning around 290 °C in oxygen and around 369 °C in nitrogen. The authors associated the faster decomposition in air with oxidation of the organic components. These values describe matrix decomposition, rather than a qualified operating range for the coated sensor.

A 532 nm laser was coupled into the coated fiber, and the emitted PL returned through the same fiber. A dichroic mirror with a 550 nm cutoff separated the excitation light from the PL. The fiber sat between the pole shoes of an iron-core electromagnet during the measurement.

A Hirst Magnetics GM08 measured the field simultaneously with the PL acquisition. The authors integrated the PL signal over 630 to 790 nm for the magnetic-field response. The spectral shape remained unchanged as the field increased, while the overall intensity decreased.

The authors attribute this response to off-axis spin-state mixing. The weak-PL state with (m_s = \pm 1) mixes with the strong-PL state with (m_s = 0) as the transverse field increases.

A single NV center has an angular response and no sensitivity to a field parallel to its crystallographic axis. The random nanodiamond orientations in the sol-gel matrix average this anisotropy. The resulting ensemble response is therefore independent of the applied field direction.

Because the readout follows NV spin-state photoluminescence, the authors expect no significant sensor hysteresis. They did not report a separate hysteresis measurement.

Only one fiber-based magnetic sensor was fabricated. Manual drop casting introduced topographic variation, so device-to-device reproducibility was not measured directly. Flat-substrate nanodiamond coatings made with several sol batches produced consistent performance, which the authors use as evidence for reproducible relative magnetic-field dependence.

A dedicated aging test was not performed on the final coated fiber. Related nanodiamond-filled layers reportedly maintained stable performance for around three years in a standard laboratory environment. Separate tests of the unfilled D50 matrix found no degradation after outdoor weathering and continuous immersion in brackish water for over four months. The matrix showed slight swelling while retaining adhesion.

The cleaved fiber tip currently limits mechanical stability. The authors propose flat-cleaved fibers with intact cladding to improve durability and coating adhesion.

Systematic temperature studies remain pending. Literature cited by the authors indicates that NV optical transitions should change little between 100 and 300 K. At higher temperatures, non-radiative recombination may affect PL intensity and magnetic contrast, so calibration over the intended temperature range may be required.

Improving thermal durability would require a lower proportion of organic components. The authors note that this change could reduce coating flexibility.

The authors propose using the sensor for magnetic current imaging in power electronics, semiconductor devices, and integrated circuits. They also identify remote monitoring inside battery packs and areas with high electromagnetic interference as possible applications. These uses were suggested from the measured response rather than demonstrated in an operating application.

The optical fiber separates the sensing tip from the readout electronics. This arrangement supports galvanic isolation, remote placement, and operation without a microwave source. The non-hysteretic response expected from the NV spin states also makes the coating compatible with static-field inspection.

Calibration must account for the changing slope. Specific sensitivity peaks between 10 and 20 mT, then decreases as the field rises. The PL signal saturates after 50 mT within the reported test range. Lock-in detection was proposed as a way to reduce PL noise and improve the detectable field contrast.

For a design based on this coating, 25 wt% nanodiamonds in D50 on a 105 μm multimode fiber tip is the demonstrated balance between PL signal, material consumption, and visible coating integrity.

View Original Abstract

Abstract The combination of sol-gel technology with nitrogen vacancy (NV - ) center-containing nanodiamonds (ND) offers a promising platform for the development of thin, functional coatings with magnetically sensitive optical properties. When applied particularly on optical fibers, these hybrid materials enable the creation of compact and highly sensitive magnetic field sensors suitable for both remote sensing and in situ applications. The spatial separation of sensing fiber tip and readout units allows flexible deployment in various magnetic field sensing applications. The developed magnetic field sensor is based on the combination of a stable, chemically resistant sol-gel layer and an embedded ND network with high magnetic sensitivity. In order to achieve homogeneous fixation of the nanodiamonds on the optical fiber, a tailored sol-gel formulation was developed, in which dimethyldiethoxysilane was selectively incorporated to ensure the required flexibility for fiber optic use. The sensor’s performance was evaluated over an operating range of 0-90 mT. It exhibits its greatest sensitivity in the crucial range up to 50 mT, demonstrated by a distinct 15% decrease in photoluminescence intensity in response to the applied magnetic field, after which the signal begins to saturate.