The uneven encapsulation caused one peak to split into two. Can this sensor still be used?
In its ideal state, the peak reflection wavelength of a Fiber Bragg Grating (FBG) adheres to the Bragg condition:
\lambda_B = 2 n_{\text{eff}} \Lambda
where n_{\text{eff}} is the effective refractive index of the core’s fundamental mode, and \Lambda is the grating period. Under isotropic conditions and uniform axial strain, the two orthogonal polarization states (x-axis and y-axis) of a single-mode fiber core are degenerate ( n_x = n_y = n_{\text{eff}} ), resulting in a single, symmetrical reflection peak.
When non-uniform packaging leads to a “split” reflection peak, it is typically caused by the following physical mechanisms. Its engineering usability must be assessed based on the specific demodulation method and application scenario:
I. Physical Mechanisms Causing Peak Splitting
-
Stress-Induced Birefringence from Transverse Non-uniform Stress
- Mechanism: Asymmetrical transverse strain difference ( \sigma_x \neq \sigma_y ) is applied to the grating core due to uneven shrinkage during packaging cure, eccentric adhesives, or asymmetric clamping by mechanical structures. Through the Photoelastic Effect, the effective refractive indices for the two orthogonal polarization modes diverge ( n_x \neq n_y ).
- Manifestation: The reflection peak splits into two polarization-dependent sub-peaks, corresponding to \lambda_{B,x} = 2 n_x \Lambda and \lambda_{B,y} = 2 n_y \Lambda . The separation between these peaks is proportional to the transverse stress difference.
-
Axial Strain Gradient within the Grating Region (Strain Gradient / Chirp Effect)
- Mechanism: Within the grating region’s length (e.g., 5\text{ mm} \sim 10\text{ mm} ), the adhesive or transfer stiffness of the packaging material is non-uniform along the axis. This causes local tension/compression in parts of the grating while other parts remain relaxed or in a different strain state.
- Manifestation: This is equivalent to cascading two gratings with different effective periods ( \Lambda_1 and \lambda_2 ), causing the original single reflection peak to split into two distinct peaks or exhibit severe asymmetrical broadening with side lobes.
II. Can a Sensor Exhibiting Peak Splitting Still Be Used?
1. Under Standard Engineering Applications and Demodulators: Generally Not Recommended (Considered Failed/Defective)
- Peak Hopping and Centroid Drift: Most commercial FBG demodulators typically rely on Gaussian fitting, thresholding, or centroid algorithms to lock onto a single peak. When the reflection spectrum shows two peaks of comparable energy, external perturbations (polarization state rotation) or minor temperature fluctuations can cause drastic changes in energy distribution between the two peaks. This easily leads to the demodulator jumping between the two peaks, resulting in significant spurious wavelength shifts (potentially tens to hundreds of \text{pm} ).
- Non-linear Distortion: For split peaks caused by axial strain gradients, the non-linear strain transfer during loading invalidates calibration coefficients.
- Conclusion: In standard temperature or strain engineering monitoring, split peaks indicate severe stress concentration or biased loading, rendering the sensor unreliable for accurate engineering measurements.
2. Under Specific Algorithms or Research Scenarios: Limited Usability Exists
- Transverse Load/Strain Sensor: If the peak splitting is purely due to birefringence from transverse forces, and the demodulator supports high-resolution full-spectrum fitting:
- The magnitude of the transverse load/pressure can be inferred by extracting the peak separation \Delta \lambda = |\lambda_{B,x} - \lambda_{B,y}| .
- The average center wavelength \bar{\lambda} = \frac{\lambda_{B,x} + \lambda_{B,y}}{2} can be used to compensate for the influence of transverse stress on temperature or axial strain measurements.
- Dual Wavelength Independent Tracking: If the demodulation software can independently track multiple peaks, and the peak separation remains consistently larger than the instrument’s resolution limit throughout the measurement range without merging, the two peaks can be treated as two independent weak reflection peaks for separate calibration.
III. Summary and Recommendations
For sensors intended for routine uniaxial strain or temperature monitoring, peak splitting caused by non-uniform packaging is a typical packaging stress defect. Replacement or re-packaging is recommended. However, in research environments studying birefringence-induced load or utilizing custom spectral fitting algorithms, the spectral splitting can be leveraged for targeted applications.