Optical characterization of wavelength-shifting and scintillating-wavelength-shifting fibers

We report results of optical characterizations of new wavelength-shifting and scintillating-wavelength-shifting fibers EJ-182 and EJ-160 from Eljen Technology and compare them to the wavelength-shifting fiber BCF-91A from Saint-Gobain. The wavelength-dependence of attenuation was derived from spectral measurements confirming that the long attenuation length increases with wavelength, while short attenuation effects become less significant at longer wavelengths. The impact of the environmental refractive index was studied by immersing the EJ-160II fiber in water. Immersing the fiber in water reduced the overall light output and suppressed the short attenuation component, which can be explained by reduced light-collection efficiency due to the smaller refractive-index contrast between the fiber cladding and the surrounding medium.

💡 Research Summary

This paper presents a comprehensive optical characterization of three newly developed fibers from Eljen Technology (EJ‑182I, EJ‑160I, EJ‑160II) and a reference wavelength‑shifting (WLS) fiber from Saint‑Gobain (BCF‑91A). All fibers have a 1 mm² square cross‑section, a polystyrene core, and a PMMA cladding (0.03 mm for BCF‑91A, 0.04 mm for the Eljen fibers). The study focuses on two key aspects: (1) the wavelength‑dependent attenuation behavior of the fibers, and (2) the influence of the surrounding medium’s refractive index on light collection efficiency.

Experimental setup
Each fiber was cut to approximately 3 m and laid in a shallow spiral groove to minimize bending losses. One end was diamond‑fly‑cut, polished, and left uncoated so that emitted light was not collected; the opposite end was coupled directly to a spectrophotometer (350–800 nm, 0.21 nm resolution). A blue LED (L200CUB500) whose emission matches the TPB spectrum was used as the excitation source. The LED could be positioned at 20 discrete distances from the spectrometer (0.124 m to 2.944 m). For each distance, ten consecutive spectra (10 ms integration each) were averaged, yielding negligible statistical uncertainty; a systematic uncertainty of 3.5 % was assigned based on repeatability tests.

Emission spectra
Figure 5 shows normalized emission spectra for all fibers at the 20 distances. As the propagation distance increases, the spectra shift toward longer wavelengths and the overall intensity drops, indicating that attenuation is wavelength‑dependent.

Attenuation modeling
The integrated light intensity (450–700 nm) versus distance was fitted with a double‑exponential model:
I(x) = I_long · e^(−x/λ_long) + I_short · e^(−x/λ_short).
The short component dominates at very short distances, while the long component governs the behavior over meter‑scale lengths. Single‑exponential fits were statistically inferior (Δχ² ≫ 6 for 490 nm and 500 nm). The fitted attenuation lengths are:

• BCF‑91A: λ_long = 3.80 ± 0.11 m, λ_short = 0.13 ± 0.04 m
• EJ‑182I: λ_long = 1.55 ± 0.11 m, λ_short = 0.54 ± 0.19 m
• EJ‑160I: λ_long = 4.00 ± 0.15 m, λ_short = 0.20 ± 0.03 m
• EJ‑160II: λ_long = 2.50 ± 0.12 m, λ_short = 0.33 ± 0.04 m

Table 2 reports the relative integrated intensity over 0–1.40 m (the LEGEND‑1000 design length) and 0–3.00 m (the actual sample length). BCF‑91A serves as the reference (1.00), while EJ‑160I and EJ‑160II retain 70–83 % of that reference, indicating competitive performance for long‑baseline detectors.

Spectral attenuation lengths
To quantify wavelength dependence, the spectra were divided into 20 nm bands and fitted separately with the same double‑exponential form, yielding Λ_long(λ) and Λ_short(λ). For wavelengths below ~500 nm, both components were required; above 510 nm, the short component became statistically unnecessary (I_short/I_total < 0.03). Λ_long(λ) increases monotonically with wavelength, reaching ~10 m at 580 nm, with localized dips near 490 nm, 610 nm, and 650 nm—features previously reported for Kuraray Y‑11 fibers. Λ_short(λ) remains < 1 m for the blue region and essentially vanishes for the main emission peak, confirming that long‑wavelength photons travel much farther.

Refractive‑index study
Because total internal reflection at the core–cladding and cladding–environment interfaces governs light confinement, the authors investigated the effect of a higher surrounding refractive index by immersing EJ‑160II in water (n ≈ 1.335). The critical angle for total internal reflection increases from ~42° (air) to ~64° (water), allowing more light to leak out. In the water‑immersed configuration, the overall integrated intensity over 0–3 m dropped to ~49 % of the air case, and the short‑attenuation component was dramatically suppressed (Figure 13). This demonstrates that in environments such as liquid argon (n ≈ 1.23) or water, the effective attenuation length is dominated by the long component, and overall light yield can be substantially reduced compared with air.

Implications for detector design
For experiments like LEGEND‑1000, which plan to use 1.4 m fibers, the data suggest that BCF‑91A and EJ‑160II provide the highest relative light yield in air. However, when the fibers are surrounded by a medium with a refractive index close to that of the cladding, the light collection efficiency can drop by roughly a factor of two. Consequently, designers must consider cladding material, thickness, and surrounding medium to maximize total internal reflection. Selecting fibers with longer λ_long (e.g., EJ‑160I) and optimizing the fiber‑to‑photodetector coupling (e.g., reflective terminations) can mitigate the loss.

Conclusions

1. Attenuation in WLS and Sci‑WLS fibers is strongly wavelength‑dependent; long‑wavelength photons experience attenuation lengths up to ~10 m, while short‑wavelength photons are quickly absorbed.
2. A double‑exponential model accurately describes the distance dependence of light intensity, capturing both a short‑range “prompt” loss and a long‑range propagation component.
3. Immersing fibers in a higher‑index medium reduces overall light output by ~50 % and suppresses the short‑attenuation component, confirming the critical role of refractive‑index contrast in light confinement.
4. Among the tested fibers, EJ‑160I shows the longest λ_long (4 m) and the smallest short‑component contribution, making it a strong candidate for long‑baseline scintillation readout, while EJ‑160II offers comparable performance with a modest penalty in higher‑index environments.

Overall, the work provides a quantitative foundation for selecting and engineering WLS/Sci‑WLS fibers in next‑generation particle‑physics detectors, emphasizing the need to match fiber optical properties to the specific optical environment of the experiment.