Full-Range Ultrasensitive Fiber-Optic Hydrogen Detection via Twin Vernier Amplification and Lightweight CNN

Area of Science:

• Optoelectronics
• Chemical Sensing
• Materials Science

Background:

• Reliable hydrogen monitoring demands sensors with high sensitivity, broad dynamic range, and environmental interference resistance.
• Existing sensors often struggle with accuracy due to temperature fluctuations and humidity.
• Integrated solutions are needed for robust hydrogen detection in safety-critical applications.

Purpose of the Study:

• To develop a monolithically integrated fiber-optic hydrogen sensor with enhanced sensitivity and accuracy.
• To achieve intrinsic temperature compensation and humidity-insensitive hydrogen detection.
• To enable end-to-end hydrogen concentration inversion across a wide range.

Main Methods:

• Fabrication of a dual-Fabry-Pérot interferometer (dual-FPI) fiber-optic sensor with separate temperature reference and hydrogen-sensitive cavities.
• Integration of polydimethylsiloxane (PDMS) and Palladium-Tungsten Oxide (Pd-WO3) materials for sensing.
• Application of digital twin Vernier spectral amplification for sensitivity enhancement.
• Development of a lightweight dual-branch fusion convolutional neural network (DBF-CNN) for data analysis and hydrogen concentration inversion.

Main Results:

• Achieved a sensitivity of -62.989 nm/% and a limit of detection of 5.33 ppm for hydrogen.
• Demonstrated intrinsic temperature compensation and humidity-insensitive operation.
• Obtained an R-squared value of 0.999443 and RMSE of 0.682% for 0-100% hydrogen concentration inversion using the DBF-CNN model.

Conclusions:

• The integrated dual-FPI fiber-optic sensor offers a compact and effective solution for hydrogen safety monitoring.
• The combination of Vernier amplification and DBF-CNN addresses key challenges in sensitivity, detection range, and environmental cross-sensitivity.
• This approach provides a robust platform for advanced hydrogen sensing applications.