Characterization of Autofluorescence in Optical Fibers for NV-based Sensing Applications

Optical fibers are crucial for guiding light in various sensing applications. Especially for quantum sensors such as the nitrogen-vacancy (NV) center in diamond, they enable light control and device miniaturization. However, fluorescence and scattering within the fiber, often referred to as fiber background, autofluorescence, or autoluminescence, can overlap spectrally with the NV centers’ fluorescence, degrading the signal-to-noise ratio and thus limiting sensor sensitivity. Here, we investigate the optical spectra of standard optical fibers, considering material dependencies, physical influences, and their fluorescence scaling with excitation power and wavelength. Our results identify spectral components and fiber types with minimal unwanted background signals, guiding the selection of optimal fibers for NV-based quantum sensing.

💡 Research Summary

The paper presents a comprehensive investigation of autofluorescence (AF) and inelastic scattering in optical fibers that are commonly employed to deliver excitation light and collect fluorescence from nitrogen‑vacancy (NV) centers in diamond for quantum sensing applications. The authors systematically examine how material composition, fiber geometry, and practical physical influences (length, bending, temperature, connectors, and adhesives) affect the spectral background that overlaps with the NV emission band (≈ 600–800 nm).

Key experimental aspects:

• A set of commercially available and self‑assembled fibers (single‑mode and multimode, low‑OH and high‑OH fused‑silica cores, polymer‑coated or metal‑coated) with lengths of 2 m and 5 m were tested.
• Excitation wavelengths spanning the NV operational range (515 nm, 520 nm, 532 nm, 594 nm, 637 nm) were launched into the fibers. Laser power was kept constant for each wavelength, and the backward‑propagating light generated inside the fiber was recorded with a calibrated spectrometer.
• Calibration of coupling losses was performed using a 650 nm reference laser and a single‑mode calibration fiber, enabling quantitative comparison of AF intensities across different fiber types.

Spectral decomposition revealed two distinct contributions:

1. Wavelength‑shifting components – Raman scattering dominates, producing peaks shifted by ~34.5 nm (≈ 560 nm) and ~48 nm (≈ 570 nm) relative to a 520 nm pump. These peaks fall within the NV detection window and therefore constitute a direct source of background. Brillouin scattering, with GHz‑scale shifts, is effectively removed by the long‑pass filters (> 650 nm) used in typical NV setups.
2. Wavelength‑static components – Defect‑related luminescence from intrinsic centers in fused silica, most notably the non‑bridging oxygen hole center (NBOHC) and the non‑bridging oxygen hole (NB‑OHC). The NB‑OHC exhibits a broad emission centered at 645 nm (1.92 eV), which matches the strongest background peak observed in all fibers. Its intensity scales with excitation power, often non‑linearly at higher powers, indicating saturation‑free defect excitation.

Material dependence: Low‑OH fused‑silica cores show the smallest Raman and defect‑related background, whereas high‑OH fibers display additional OH‑related peaks around 1385 cm⁻¹, increasing overall AF by a factor of 2–3. Metal (aluminum) cladding suppresses thermally induced fluorescence compared with polymer coatings. Self‑assembled fibers using black tubing and black connector boots reduce stray reflections and exhibit 5–7 % lower background than standard yellow/orange‑sheathed commercial fibers.

Physical influences:

• Length – AF intensity grows linearly with fiber length; a 5 m fiber produces roughly 2.5× the background of a 2 m fiber.
• Bending – Radii below 5 mm introduce micro‑cracks and strain‑induced defects, raising AF by 10–20 %.
• Temperature – Raising the fiber temperature from 20 °C to 80 °C increases AF by ~80 % on average, reflecting thermally activated defect emission.
• Connectors & adhesives – Epoxy used in FC/PC connectors emits weakly (≈ 0.3 % of the NV signal) in the 600–700 nm band; this contribution becomes non‑negligible for high‑sensitivity measurements.

A quantitative comparison with NV‑diamond fluorescence (15 µm diameter, 3.5 ppm NV concentration) shows that the AF from a 2 m low‑OH SM fiber equals the NV signal when the diamond size is reduced to ≈ 740 nm. Consequently, for sub‑micron diamonds or low‑NV‑density samples, fiber background can dominate the detected signal.

Based on these findings, the authors propose practical guidelines for NV‑based fiber‑coupled sensors:

• Choose low‑OH fused‑silica, single‑mode fibers with metal cladding.
• Keep fiber length ≤ 2 m and avoid tight bends (≥ 10 mm radius).
• Operate at modest temperatures (≤ 30 °C) or implement active cooling.
• Prefer connector‑free splices or low‑fluorescence adhesives; if connectors are required, use silicone‑based or UV‑cured low‑fluorescence epoxies.
• Employ long‑pass filters with cut‑on wavelengths > 650 nm to suppress Raman peaks, and consider additional notch filters at 645 nm to attenuate NB‑OHC emission.

In summary, the paper delivers a detailed, experimentally validated map of how different fiber properties and environmental conditions contribute to autofluorescence that interferes with NV‑center readout. By following the recommended fiber selection and handling protocols, researchers can substantially improve the signal‑to‑noise ratio of quantum magnetic‑field sensors, enabling higher sensitivity and more reliable operation in compact, fiber‑integrated platforms.