Skip to content

Quantum Sensing of Magnetic and Thermal Fields Using Nitrogen-Vacancy Centers on Spoof Surface Plasmon Waveguides

FieldValue
JournalDigital Scholarship - UNLV (University of Nevada Reno)
Published2026-05-11
AuthorsMaxwell Stonham
DOI10.34917/40245777

A 120 μm fabricated gap produced the strongest overall ODMR response in a comparison of three copper spoof surface plasmon waveguides. A focused retest near the gap floor reached 18% to 19% contrast on the 285 μm structure. The same hybrid device detected magnetic fields and tracked temperature, with an average ODMR frequency slope of -175 kHz/K.

COMSOL models used the actual fabricated gaps rather than their design labels. At 2.87 GHz, the maximum simulated field norm decreased as gap width increased. The following table compares that field with the initial ODMR contrast measured in the gap at 100 mW laser power and 7 dBm RF input.

Nominal labelActual gapSimulated peak fieldInitial in-gap contrast
100 μm120 μm294 A/mAbout 10%
200 μm210 μm139 A/m10%
300 μm285 μm133 A/m12%

The 285 μm device produced slightly more contrast than both narrower structures. The authors linked this exception to focus position and sample uniformity. The simulated microstrip maximum was 33.7 A/m, compared with 4 to 6 A/m through its center. Within each gap, the field was strongest at the bottom and weakened toward the top.

Retesting near the gap floor gave 15% contrast for the 210 μm structure and 18% to 19% for the 285 μm structure. These tests used 50 mW laser power and RF sweeps from 2.7 GHz to 3 GHz. Placement could therefore outweigh nominal gap width in a single measurement, while the simulated field still increased as the gap narrowed.

Prior work had experimentally characterized a 2 mm gap, while validation across smaller gaps remained incomplete. The present work compared actual gaps of 120 μm, 210 μm, and 285 μm and added thermal testing.

NV centers are atomic-scale defects formed by adjacent nitrogen and vacancy in diamond. The negative NV state is the magneto-optically active state used for sensing. A 532 nm laser excites these centers, and their fluorescence spans 630 nm to 800 nm. Microwave transitions between the ms = 0 and ms = ±1 spin states produce ODMR dips.

At room temperature, the ground state zero field splitting is 2.87 GHz. A magnetic field separates the ms = −1 and ms = +1 resonances through the Zeeman effect. Temperature shifts the splitting at about -74 kHz/K.

The copper SSPW uses periodic rectangular gaps to support a surface bound TM mode. Its concentrated microwave field drives the NV transitions. The reported experimental outcomes were ODMR contrast, magnetic splitting, and thermal frequency shifts. An optical emission gain was not reported as a separate measurement.

Each device used Rogers RT/duroid 6010.2LM, a 127 mm square ceramic PTFE laminate with 17.5 μm copper on both sides. The structures measured about 13 cm long, 3 cm high, and 0.64 mm thick. The substrate had a relative permittivity of 10.2 and a loss tangent of 0.0023.

Photoresist AZ P4110 was spun at 2000 RPM, with 10 seconds of acceleration, for 35 seconds total. The coated substrate was baked at 100°C for 1 minute 15 seconds. Each maskless exposure window lasted 160 seconds. Development took about 25 seconds, followed by ferric chloride etching for about 30 minutes at 24°C.

The nanodiamond solution contained about 2.5 ppm NV centers and had an average particle size of 70 nm. Two drops taken from a 5 mL, 1 mg/mL solution in deionized water were air dried over the gap and microstrip.

A 532 nm laser operated at 10 mW to 200 mW and focused through a 40× objective with 0.65 NA and 0.6 mm working distance. Fluorescence was recorded by a spectrometer and processed in MATLAB. RF swept from 2.5 GHz to 3 GHz through a 16 W amplifier. A 50 Ω, 25 W, 3 GHz dummy load terminated the other waveguide port. A vector network analyzer supplied S11 and S21 measurements.

At the initial test condition, the 120 μm structure produced about 10% contrast in the gap, 6% near the gap, 6% on the microstrip, and 2% between gaps. A later microstrip rerun reached 8% to 10%, showing that focus position strongly affected the result.

During a 7 dBm RF sweep, the 120 μm gap reached 10% contrast at 20 mW, 30 mW, and 40 mW laser power. It reached 8% at 10 mW, 50 mW, and 60 mW. The authors then used 40 mW for the RF comparisons.

At 10 mW laser power and -18 dBm RF input, the estimated output after amplification was 27 dBm, or 0.50 W. The contrast was about 5.5%, and the Zeeman splitting became visible. Lower drive narrowed the resonance, while strong microwave fields caused power broadening.

A 6 mm cube magnet supplied the external field, and a gaussmeter provided an independent magnitude check. The 285 μm comparison table reports 0.41 mT from ODMR and 0.64 mT from the gaussmeter at the 35 mm position. At 15 mm, the values were 1.57 mT and 4.48 mT. The resonance separations increased from 23 MHz to 88 MHz.

ODMR senses the field projected onto each NV axis. Random NV orientations make the gaussmeter magnitude differ from the ODMR projection. The authors identify the most aligned centers as the responding subset.

Stronger fields widened the splitting and reduced contrast as the population divided between two transitions. On the 120 μm device at 10 mW laser power and -18 dBm input, contrast fell from 2.5% at 25 mm to 1.5% at 20 mm.

Thermometry was tested only on the 120 μm structure. TEC setpoints ran from 20°C to 35°C in 1°C steps, with 7 dBm RF input and a 2.84 GHz to 2.89 GHz sweep at 75 kHz steps. Near gap broad sweeps at 50 mW to 150 mW produced fitted slopes of -147 to -200 kHz/K. A 24°C to 25°C sweep at 50 mW gave -176 kHz/K. An in gap sweep at 30 mW gave -196 kHz/K.

The authors report an average slope of -175 kHz/K, 2.36 times the known -74 kHz/K coefficient. They attribute the difference to localized heating and imperfect thermal contact with the feedback thermistor. The thermistor sat on the front dielectric rather than at the NV centers.

Each 1°C TEC increase produced about 2.2°C at the NV centers on average. At a fixed 25°C TEC setpoint, raising the laser from 10 mW to 100 mW changed the inferred temperature from 43.2°C to 66.1°C. The fitted slope was -18.5 kHz/mW, equivalent to 0.25 K per mW.

The authors suggest that the nearly linear shift can support calibrated thermometry. They do not present the -175 kHz/K value as an intrinsic diamond coefficient.

Fabrication control was a major limitation. The fabrication record showed roughly 60 μm to 100 μm more spacing than some lithography targets. Warming ferric chloride to 40°C to 45°C reduced a test etch from 30 minutes to 15 minutes and produced 50 μm to 60 μm gaps on one segment. No complete waveguide followed that change.

The 0.6 mm objective working distance also made laser placement hard to verify. The 285 μm exception likely reflects small focus or sample differences. Thermal testing had one waveguide, a fragile assembly, fan vibration, and controller fluctuations of tens of mK. These effects increased scatter in 0.1°C sweeps.

For device engineering, smaller gaps increased the simulated field and reduced the optical and RF power needed for strong ODMR contrast. Engineers still need to measure the etched gap and register the laser to the high field region. The tested diamond was a commercial 70 nm NV nanodiamond ensemble on copper clad Rogers 6010, so the data do not establish performance for bulk single crystal diamond. A calibrated assembly must report the actual gap, focus location, magnetic projection, and thermal transfer before its readings are used as field or temperature measurements.

View Original Abstract

Nitrogen-vacancy (NV) centers in nanodiamonds are solid-state quantum defects capable of detecting magnetic fields and temperature at the nanoscale with high spatial resolution and optical addressability at room temperature. Their spin-dependent fluorescence enables optically detected magnetic resonance (ODMR), a technique widely used for quantum sensing. However, enhancing optical readout efficiency remains a critical challenge for scalable integration and various methods can be used to improve this. This thesis explores a hybrid quantum sensing platform that integrates NV-based sensing with spoof surface plasmon waveguides designed to enhance optical excitation and emission efficiency through localized spoof plasmon modes supported by subwavelength corrugated gaps. A combination of physics-based simulations and experimental fabrication is used to characterize the interactions between NV center fluorescence and spoof surface plasmon waveguides of varying gap sizes. This work demonstrates dual-mode magnetic and thermal field sensing within a compact hybrid system by evaluating how spoof plasmonic structures influence fluorescence contrast, magnetic resonance response, and temperature-dependent spectral shifts. These results aim to contribute towards the development of scalable, high-sensitivity photonic quantum sensors used for nanoscale sensing in microelectronics, quantum materials, and biomedical devices.