Abstracts:A deep learning framework-based BiLSTM-assisted on-chip Fabry-Perot (FP) sensor for high precision dual-parameter prediction is demonstrated, in which the open sensing cavity functions as a sensing window and is directly exposed to the analyte solution. Variations in the glucose concentration and temperature of the solvent within the sensing cavity induce corresponding wavelength shifts in the interference spectrum, enabling the simultaneous measurement of the glucose concentration and temperature for the demonstration. Experimental results show that the proposed sensor achieves a sensitivity of 0.97 nm/(g/L) (equivalent to 1070 nm/RIU), and a temperature sensitivity of 33 pm/°C. To overcome the low efficiency and noise susceptibility of conventional interferometric spectrum demodulation methods in multi-parameter and large-volume data scenarios, a BiLSTM network is employed to perform high-precision demodulation of the experimentally acquired spectral sequence data regardless of the wavelength and the free spectral range. Through bidirectional sequence modeling, the BiLSTM effectively captures both forward and backward correlation information embedded in the spectral sequences and intensity, enabling accurate discrimination of temperature- and concentration-induced spectral variations with a coefficient of determination (R2) exceeding 0.99 and a root mean square error (RMSE) as low as 1.5134 × 10−2 achieved. The proposed dual-parameter FP-based on-chip sensor features a compact configuration and exhibits high stability and repeatability, which provides a viable approach for the high-precision simultaneous detection of glucose concentration and temperature, and offers a valuable reference for the integration of deep learning techniques with on-chip interferometric sensing platforms.

Abstracts:Spun quarter-wave plates (QWPs) can be co-fabricated and packaged with spun fiber sensing coils, which reduces fusion splices and improves device consistency. Their distributed birefringence and extended length also increase sensitivity to mechanical vibration. In reflective all-fiber optic current sensors (FOCS), vibration can change the interferometric visibility. This change acts as a time-varying equivalent closed-loop gain and can affect AC current demodulation. We develop an analytical polarization transfer model for a spun QWP and describe vibration as a time-varying perturbation of the dimensionless rotation parameter. We then propose an eight-state square-wave modulation with synchronous three-channel demodulation. Within one modulation period, it yields a current error signal, a half-wave-voltage drift observable, and a visibility observable. The visibility-related channel can be used to track the vibration-induced waveform. It can also be used to normalize the current error so that the effective loop gain is less sensitive to visibility changes. Tests from 0 to 30 Hz indicate that, under these low-frequency conditions, the sensing-coil phase disturbance is small in the reflective configuration. The dominant effect is visibility variation associated with the spun QWP. With the proposed normalization, the mean standard deviation of the extracted AC-current peaks and valleys decreases from 0.27 A to 0.094 A at 50 Hz and from 0.28 A to 0.089 A at 100 Hz. The method does not require additional components and can improve demodulation robustness in vibratory environments.

Abstracts:The application of high-quality-factor (high-Q) and multi-mode whispering gallery mode (WGM) microcavities significantly enhances optical power utilization and rotation responsivity in broadband light source-driven resonant micro-optic gyroscopes (RMOGs). However, interactions between different transverse modes within the WGM microcavity raise the baseline level of relative intensity noise (RIN), thereby limiting the practical sensitivity of the gyroscope system. This paper first establishes the feasibility of the equivalent mode method under multi-mode detection. It then demonstrates that the RIN of broadband light after transmission through a WGM microcavity consists of two components: one arising from the independent response of each mode, and the other resulting from inter-mode interactions. To analyze the latter, a uniform distribution hypothesis for the resonant frequency differences between modes is proposed. Based on this hypothesis, it is demonstrated that a microcavity with high-Q, high transmittance, and a modest—though not single—number of modes can minimize the total noise in the gyroscope system. Finally, a CaF2 crystal WGM microcavity with an effective number of 20 transverse modes, an equivalent Q-factor of 9.8×108, and a diameter of 28.5 mm was fabricated. Utilizing this microcavity as the core sensing element, an RMOG was constructed, achieving an angle random walk of 0.095 °/√h and a bias instability of 0.17 °/h.

Abstracts:Integrating highly ordered nanostructures onto the curved surfaces of optical fibers remains a significant fabrication challenge that limits the application of localized surface plasmon resonance (LSPR) in fiber sensing. To address this, we developed a flexible polyvinyl alcohol (PVA)-assisted transfer strategy that enables the conformal integration of large-area, periodic silver nanodisk arrays (SDAs) onto multimode fibers. Leveraging this technique, we constructed a hybrid plasmonic architecture by functionalizing the SDAs with Ti3C2Tx-MXene nanosheets. Theoretical and experimental results indicate that this composite structure excites a hybrid mode through plasmon hybridization between the LSPR at the nanodisk edges and the propagating SPR supported by an auxiliary silver layer. Additionally, the MXene overlayer improves sensing performance via its high conductivity and molecular enrichment capabilities. The resulting sensor achieves a refractive index sensitivity of 3336 nm/RIU in the 1.335–1.375 range and exhibits versatile sensing capabilities, functioning effectively for RI detection in liquids and, upon further PVA encapsulation, as a highly sensitive humidity monitor with a humidity sensitivity of 0.928 nm/%RH. This study provides a reliable, cost-effective route for integrating ordered micropatterns onto non-planar optical substrates.

Abstracts:Accurate cryogenic temperature sensing is critical for superconductivity, space technology, and low-temperature biomedicine. We present a high-sensitivity fiber-optic cryogenic temperature sensor based on Vernier effect amplified cascaded Sagnac interferometers (CSIs). The sensing arm incorporates a nematic liquid crystal (NLC)-embedded Sagnac interferometer (SI), while the reference arm uses a polarization-maintaining fiber (PMF)-based SI with a slightly detuned free spectral range (FSR), generating the Vernier effect for sensitivity enhancement. A fiber Bragg grating (FBG) with monotonic linear cryogenic response is integrated into the sensing loop to provide wavelength markers for coarse temperature estimation and FSR period identification, overcoming spectral aliasing and extending the measurement range. Experiments show an average sensitivity of 14.43 nm/K over 260–210 K, a 6.36-fold improvement over a single SI. These results validate NLCs as effective cryogenic thermometric materials and demonstrate the sensor’s potential for applications such as short term biobanking of temperature-sensitive biomaterials and low temperature mechanical testing in industrial environments.

Abstracts:We present a spectroscopic method using a silicon-photonic-integrated laser by frequency-modulated continuous-wave (FMCW) interferometry to realize rapid multipoint gas sensing. The employed hybrid integrated external cavity laser has a wide tuning range of 147 GHz, a narrow linewidth of 180 kHz and a high sweeping speed of 294 THz/s (∼2700 nm/s), enabling sub-millisecond multipoint spectroscopic gas sensing. In the proof-of-concept experimental demonstration, three methane sensing nodes are deployed along a 30-m optical fiber, achieving a measurement sensitivity of 6 ppm with a time resolution of 0.5 ms and a positioning spatial resolution of 0.7 mm. Having the advantage of high spatiotemporal resolution, our proposed method promotes a novel way of developing spectroscopic gas sensors for challenging safety monitoring applications where spatially resolved chemical analysis with fast response is required.

Abstracts:The conflict between the array element spacing and the localization accuracy has always been a difficulty for the sound source localization array. The coupled audio systems of small living creatures inspire the miniaturized acoustic directional sensors with interaural difference amplification to resolve the conflict. However, the strong resonance characteristics of the coupled system make it only work in a narrow resonance zone. In this paper, we proposed an optical fiber Fabry-Perot (F-P) directional acoustic sensor based on an asymmetric coupled cavity, whose frequency range, with the interaural differences amplification, breaks the resonance limitations. The proposed coupled cavity system is simplified to a two-degree-of-freedom (2-DOF) lumped mechanical model. Based on the electro-mechano-acoustical (EMA) analogy, the dynamic coupling between the membrane and cavity pressure field is explained by the equivalent circuit model. Numerical simulations based on the lumped model indicated that the interaural differences are related to the azimuth angles and are effectively amplified in a wider frequency range than the symmetric model. An optical fiber F-P cavity-coupled directional acoustic sensor with an element spacing of 6 mm was fabricated, and the directional experiments of the sensor were performed from 2 kHz to 3 kHz. The width of the resonance zone with the amplified interaural differences is 740 Hz, consistent with the proposed theoretical model. The directional sensitivities based on the interaural intensity differences are in the range of 2.14 nm/(Pa·°)–13.95 nm/(Pa·°), and the angular resolution is 1.50°–3.75° in the range from −60°–60°.

Abstracts:Mid-infrared fiber evanescent wave spectroscopy offers a promising route for gas sensing due to its molecular specificity. This study demonstrates an enhanced mid-infrared gas sensor for SF6 detection at 10.55 μm, based on a metal-organic framework (MOF) coated chalcogenide taper-in-taper fiber (TITF). The TITF, fabricated from a custom-drawn Ge-As-Se-Te fiber, features primary and secondary waist diameters of 20 μm and 10 μm, respectively. This structure significantly enhances evanescent field leakage, yielding a 113% sensitivity improvement over a conventional tapered fiber. Furthermore, coating the waist with zeolitic imidazolate framework-8 (ZIF-8) concentrates SF6 molecules at the fiber surface via gas adsorption, providing an additional 88% sensitivity enhancement. To the best of our knowledge, this is the first report that integrates a TITF structure with MOFs for infrared optical fiber gas sensing. This work provides an effective strategy combining geometric optimization and functional coating for developing high performance optical fiber gas sensors.

Abstracts:High-resolution manometry (HRM), as a critical diagnostic method for gastrointestinal diseases, is widely employed for esophageal manometry. However, existing water-perfused catheters and expensive solid-state catheters to realize HRM suffer from either limited frequency response or insufficient spatial resolution. On the other hand, fiber-optic manometry sensor based on optical frequency domain reflectometry (OFDR) can realize distributed measurement at high spatial resolution. Nevertheless, there is a tradeoff between strain resolution and sensing resolution in conventional OFDR, which prevents it from achieving both high sensing resolution and pressure resolution in esophageal manometry, making it difficult to satisfy the HRM requirement. To address this, we propose a spectrogram reassignment (SR) method for strain demodulation, which simultaneously enhances the sensing resolution and strain resolution in OFDR but without much deterioration of the demodulation speed. The experiment results demonstrate a sensing resolution of 1.4 mm and strain resolution of 1.9 μϵ, showing over 4 times improvement when compared with the conventional OFDR. Using the fabricated fiber-optic manometry sensor with 80 μm diameter bend-insensitive fiber, our SR-assisted OFDR demonstrates a pressure sensitivity of 1.97 μϵ/mmHg and pressure resolution of 0.96 mmHg within traditional esophageal pressure range of 0∼300 mmHg, satisfying the HRM requirement. This is the first work to present an OFDR fiber-optic manometry sensor that meets the HRM requirement, and it provides a competitive method for esophageal manometry.

Abstracts:Optical microcombs with broad spectral coverage are critical for applications such as optical communications, spectroscopy, and metrology. This is typically achieved by introducing dispersive waves to enhance the comb power at spectral edges. In silica microspherical cavities, dispersion engineering is strongly coupled to its diameter, leaving a challenge to manage wide-range and high-order dispersion while maintaining high-quality factors suitable for efficient microcomb generation. Here, we fabricate silica microspheres with precisely controlled diameters and quality factors exceeding $10^{8}$. By targeting specific diameters to position the zero-dispersion point near the pump, we realize dispersive wave-enhanced microcombs in both anomalous and normal dispersion regimes. In the anomalous dispersion regime, Kerr combs with a 360 nm bandwidth are generated. In the normal dispersion regime, we achieve continuous hybrid Raman–Kerr combs spanning 380 nm and observe the spectral evolution from a Raman–Kerr state to a Kerr-dominated comb. Numerical modeling elucidates and reproduces the underlying dynamics under both dispersion conditions. Furthermore, the simulations predict the feasibility of accessing stable dissipative Kerr solitons and mode crossing-assisted dark pulses. Our results provide insights into dispersive wave dynamics across distinct dispersion regimes, highlighting the potential for the flexible generation of broadband, coherent microcombs.