Abstract

Long-term online monitoring in extreme environments (e.g., high temperature, high pressure, liquid corrosion, strong electromagnetic radiation, confined narrow spaces, etc.) poses severe challenges to the stability and reliability of sensors. Optical fiber sensors, which employ optical signals as carriers, leverage inherent advantages such as immunity to electromagnetic interference, corrosion resistance, and electrical insulation, making them a critical technological route for extreme environment sensing. Among these, fiber-optic Fabry–Perot (F–P) sensors, operating by measuring changes in the interference cavity length, feature a simple structure, ease of miniaturization, high sensitivity, and high dynamic response. They are capable of detecting nanoscale displacement, vibration, and other minute signals and can be extended to measure multiple parameters (e.g., temperature, pressure, acceleration, etc.) via external sensitive structures. Consequently, they have become a research hotspot in high-precision optical fiber sensing. This paper systematically reviews fiber-optic F–P sensing technology. It first elaborates on the multi-beam interference mechanism and cavity length demodulation methods. It then analyzes the development history of sensing structure designs for physical quantities such as displacement, temperature, pressure, acceleration, and vibration, with particular emphasis on packaging techniques and demodulation strategies suitable for high temperature, high pressure, liquid media, and strong electromagnetic environments. This review introduces the engineering practices of such sensors in aerospace, oil and gas extraction, nuclear reactor monitoring, and other related fields. Finally, it discusses the progress in developing new-material optical fibers (e.g., sapphire fiber) and integrated probe assemblies suited for more extreme environments, such as those involving ultra-high temperature and strong irradiation. It explores the feasibility of applying artificial intelligence to adaptive decoupling under multi-field coupling, as well as miniaturized networking schemes based on integrated on-chip spectrometers. Furthermore, it envisions a trend toward the large-scale deployment of standardized and engineered measurement systems. This work provides a systematic reference for in-depth research and engineering applications of fiber-optic F-P sensing technology in extreme environments.

Keywords

fiber-optic fabry–perot (F–P) sensor; spectral demodulation; multi-field coupling; extreme environment; decoupling

1 Introduction

In practical engineering measurements, sensors that are long-term exposed to coupled extreme environments, such as high temperature, high pressure, corrosive liquid, strong electromagnetic interference, and intense radiation, while being required to operate stably in confined and narrow spaces, impose stringent demands on measurement stability, reliability, and long-term operational capability^1–3. Conventional electrical sensors, which rely on metallic wires and electronic components, are prone to insulation degradation, electromagnetic induction noise, and other forms of signal crosstalk in high-temperature, humid, corrosive, and strong electromagnetic environments. In severe cases, these issues can lead to device burnout, making it difficult to meet the requirements of long-term online monitoring^4–6. However, with the advancement of intelligent major equipment and extreme engineering environments, deep-earth energy extraction requires long-term in-situ monitoring of temperature, pressure, strain, vibration, and other parameters; deep-sea equipment demands stable operation under high hydrostatic pressure and strong corrosion; and nuclear power systems require high radiation resistance and high safety with long-term maintenance-free capability^7–9. Therefore, the development of novel sensing technologies for extreme environments has become an important research direction in this field.

Compared with conventional electrical sensors, optical fiber sensors, which use optical signals as carriers, offer advantages such as immunity to electromagnetic interference and corrosion resistance, making them more suitable for extreme working conditions including high temperature, high pressure, liquid corrosion, and intense radiation^10,11. According to their operating principles, optical fiber sensing technologies are mainly classified into two types: distributed and point-based. Distributed sensors are applicable to large-scale measurement scenarios such as bridge safety monitoring, whereas point-based sensors are more focused on local physical quantity detection¹². Since their introduction, point-based optical fiber sensing technologies have evolved from intensity-based types toward interferometric and grating-based types capable of multi-parameter fusion sensing. Fiber Bragg gratings and fiber-optic Fabry–Perot (F–P) sensors have gradually become mainstream devices, collectively promoting the development of collaborative sensing of multiple parameters with high sensitivity and high stability¹³. As monitoring demand in increasingly complex environments grow, optical fiber sensing technology is transitioning from single-parameter measurement toward multi-parameter fusion for long-term online monitoring, thereby gradually establishing itself as a key technical approach for extreme environment sensing.

After years of development, point-based optical fiber sensing technologies have formed multiple technical routes, represented by F–P, fiber Bragg grating (FBG), fiber evanescent wave, and fiber surface plasmon resonance (SPR)¹⁴. Different sensing mechanisms and structural characteristics make them suitable for different application scenarios, collectively promoting the development of complex environment sensing. FBG sensors, by virtue of wavelength-encoded stability, immunity to light intensity fluctuations, and wavelength division multiplexing capability, are widely used in large-scale networked measurements. However, limited by the principle of refractive index modulation in the fiber core, their sensitivity is constrained by wavelength drift, resulting in relatively limited capability for detecting minute signals. Moreover, the decoupling of temperature and strain is complex¹⁵. Fiber evanescent wave sensors utilize the evanescent field that leaks into the cladding to sensitively detect changes in external refractive index and liquid media. Fiber SPR sensors further incorporate the surface plasmon resonance effect of metal thin films, exhibiting outstanding performance in chemical applications such as biomedicine, but they rely heavily on the external dielectric environment^16–18. In contrast, fiber-optic F–P sensors operate based on changes in the interferometric cavity length, featuring a simple structure, ease of miniaturization, and both high sensitivity and high dynamic response. Their optical path length sensitive mechanism enables nanoscale monitoring of tiny displacement and vibration. Moreover, by introducing external sensitive structures, extrinsic sensors for temperature, pressure, acceleration, and other parameters can be constructed^19,20. The four typical point-based optical fiber sensors mentioned above are shown in Fig. 1. Among them, the fiber-optic F–P sensor, by virtue of its unique interferometric measurement mechanism and excellent structural designability, has become a research hotspot in high-precision optical fiber sensing. The F–P sensor constructs a miniature optical resonant cavity to convert changes in external physical quantities into optical phase changes or interference spectral shifts, thereby enabling high-sensitivity detection of parameters such as temperature, pressure, strain, displacement, vibration, and acoustic signals (as illustrated in Fig. 2).

Figure 1. Main categories of fiber-optic point sensors.

Figure 2. Sensing mechanism of fiber-optic F–P sensors. (a) Schematic of a composite F–P cavity consisting of the inner cavity FP1, outer cavity FP2, and composite cavity FP3. (b) Sensing mechanisms of the composite F–P cavity for different measurands, where n denotes the refractive index of the measurement environment.

Existing reviews have systematically surveyed fiber-optic Fabry–Perot sensing technologies from various perspectives. Islam et al. reviewed the development history, fabrication methods, operating principles, and applications of fiber-optic F–P sensors²¹ while Huang et al. focused on F–P interference models and different cavity configurations, summarizing their applications in refractive-index, temperature, and micro-displacement sensing²². As F–P sensors are increasingly extended toward extreme environments involving high temperature, high pressure, corrosive media, and intense radiation, research interests have broadened to include extreme-environment-resistant materials and packaging, multiphysics coupling, multiparameter decoupling, and high-precision intelligent demodulation. Accordingly, this review takes extreme-environment monitoring as its central theme and integrates sensing structures and materials, multiphysics coupling, multiparameter measurement, demodulation techniques, environmental adaptability, and engineering applications into a unified framework.

This paper presents a systematic review of fiber-optic F–P sensing technology. It begins by elucidating the core measurement mechanism and then traces the evolution of sensing structure designs for various physical quantities, including displacement, gap, temperature, pressure, acceleration, and vibration. Particular emphasis is placed on F–P sensing structures suited for complex environments characterized by high temperature, high pressure, liquid media, and strong electromagnetic fields. The paper systematically introduces packaging processes, demodulation techniques, and scenario-tailored engineering solutions. Furthermore, it discusses recent advances in the development of novel-material optical fibers and integrated probe assemblies for more extreme conditions, and evaluates the feasibility of multi-field coupled adaptive decoupling algorithms combined with artificial intelligence approaches, as well as miniaturized networking schemes based on integrated on-chip spectrometers. Finally, the paper outlooks the trend toward large-scale deployment of standardized and engineered measurement systems.

2 Core fundamental sensing mechanism and multi-field coupled response theory of fiber-optic F–P sensors

2.1 Fundamental theory of fiber-optic F–P interferometric sensing

Fiber-optic F–P sensors are typical interferometric fiber-optic sensors, and their core mechanism is based on multiple reflections and coherent superposition of light waves within an optical resonant cavity. Compared with intensity-modulated sensors, F–P sensors achieve physical quantity sensing by detecting optical phase changes, offering higher measurement resolution and sensitivity.

An F-P interference cavity consists of two parallel reflective surfaces that define an optical cavity of length L. Based on the structural configuration, the F–P cavity can be classified into two types: an intrinsic F–P interferometer (IFPI) and an extrinsic F–P interferometer (EFPI). The intrinsic type is formed directly within the optical fiber or through a sealed fabrication process, making it suitable for applications involving minute deformations induced by pressure, temperature, or other physical parameters. The extrinsic type creates an air cavity between the fiber end face and an external reflective surface, offering high structural flexibility and ease of packaging.

(1) Mechanism of F–P interference formation:

Let the incident electric field be:

{E}_{i}={E}_{0}{e}^{j\omega t}\;,\[{E}_{i}={E}_{0}{e}^{j\omega t}\;,\] (1)

where E₀ is the amplitude of the incident light; ω is the angular frequency; t is time.

When the light wave reaches the first reflective surface, a portion of the light is reflected, generating the first reflected light beam E₁. Another portion enters the F-P cavity, is reflected by the second reflective surface, and returns to form the second reflected light beam E₂. The electric fields of the first and second reflected light beams are:

{E}_{1}={r}_{1}{E}_{0}{e}^{j\omega t} \;,\[{E}_{1}={r}_{1}{E}_{0}{e}^{j\omega t} \;,\] (2)

{E}_{2}={t}_{1}{t}_{2}{r}_{2}{E}_{0}{e}^{j(\omega t+\phi )} \;,\[{E}_{2}={t}_{1}{t}_{2}{r}_{2}{E}_{0}{e}^{j(\omega t+\phi )} \;,\] (3)

where r₁ and r₂ are the reflection coefficients; t₁ and t₂ are the transmission coefficients; Φ is the phase difference between the two beams.

Since the second reflected light beam propagates back and forth once inside the F-P cavity, the optical path difference between the two beams is:

\Delta =2nL\;,\[\Delta =2nL\;,\] (4)

where n is the refractive index of the medium inside the cavity; L is the F-P cavity length.

The phase difference corresponding to the optical path difference is:

\phi=\frac{2\pi}{\lambda}\Delta=\frac{4\pi nL}{\lambda}\; .\[\phi=\frac{2\pi}{\lambda}\Delta=\frac{4\pi nL}{\lambda}\; .\] (5)

Due to the principle of interference superposition, the total output light intensity is:

I={\left| {E}_{1}+{E}_{2}\right| }^{2}={I}_{1}+{I}_{2}+2\sqrt{{I}_{1}{I}_{2}}\cos \phi \;,\[I={\left| {E}_{1}+{E}_{2}\right| }^{2}={I}_{1}+{I}_{2}+2\sqrt{{I}_{1}{I}_{2}}\cos \phi \;,\] (6)

where I₁=|E₁|², I₂=|E₂|².

When Φ=2mπ, constructive interference occurs, and the output light intensity reaches its maximum. When Φ=(2m+1)π, destructive interference occurs, and the output light intensity is at its minimum. Therefore, the fiber-optic F-P sensor essentially achieves the sensing of external physical quantities by detecting the interference phase change.

(2) Multi-beam interference mechanism

For a high-reflectivity F–P cavity, multiple reflected beams are generated inside the cavity. In this case, the multi-beam interference model is adopted for analysis. Let the cavity surface reflectivity and transmissivity of each cavity surface be R and T, respectively. The transmitted light intensity follows the Airy distribution. The corresponding light intensities of the transmitted and reflected light are given by:

{I}_{t}={I}_{0}\frac{1}{1+F{\sin }^{2}(\phi /2)} \;,\[{I}_{t}={I}_{0}\frac{1}{1+F{\sin }^{2}(\phi /2)} \;,\] (7)

{I}_{r}={I}_{0}\frac{F{\sin }^{2}(\phi /2)}{1+F{\sin }^{2}(\phi /2)} \;,\[{I}_{r}={I}_{0}\frac{F{\sin }^{2}(\phi /2)}{1+F{\sin }^{2}(\phi /2)} \;,\] (8)

where F=4R/(1-R)², which is the finesse coefficient.

As the reflectivity increases, the interference spectral linewidth decreases, the spectral peaks become sharper, the interference contrast improves, and the spectral resolution increases. Consequently, the measurement system exhibits more sensitive to both the stability of the light source and the influence of external mechanical vibrations.

(3) Modulation mechanism by external physical quantities

Fiber-optic F–P sensors belong to the category of optical path length modulated sensors. Their core measurement parameter is the optical path difference. External physical quantities change the interference phase by altering the cavity length L and the refractive index n. When small changes occur in the cavity length and refractive index, the corresponding phase changes are:

\Delta {\phi }_{L}=\frac{4\pi n}{\lambda }\Delta L \;,\[\Delta {\phi }_{L}=\frac{4\pi n}{\lambda }\Delta L \;,\] (9)

\Delta\phi_n=\frac{4\pi L}{\lambda}\Delta n\; .\[\Delta\phi_n=\frac{4\pi L}{\lambda}\Delta n\; .\] (10)

The total phase change is:

\Delta\phi=\frac{4\pi}{\lambda}(L\Delta n+n\Delta L)\; .\[\Delta\phi=\frac{4\pi}{\lambda}(L\Delta n+n\Delta L)\; .\] (11)

Phase variation constitutes the theoretical basis for fiber-optic F-P sensors in sensing physical quantities such as temperature, pressure, strain, and vibration.

2.2 Common mechanism of multi-physical quantity cross-coupling interference in extreme environments

Fiber-optic F–P sensors achieve high-precision measurement based on optical phase changes and offer significant advantages under extreme operating conditions. However, their operating mechanism involves complex multi-field coupling problems. Multiple physical quantities such as temperature, pressure, strain, and vibration act simultaneously on the F–P cavity, causing coupled changes in cavity length L and refractive index n and producing cross-sensitivity. Factors such as sweep frequency nonlinearity, laser frequency jitter, packaging residual stress, and material thermal drift further degrade measurement stability and demodulation accuracy. Especially under high-temperature and high-dynamic conditions, multi-physical field coupling leads to interference spectrum broadening, spectral peak drift, and accumulation of frequency-domain demodulation errors. Essentially, any factor that alters either L or n induces a change in the interference phase. The strong nonlinearity and multi-scale characteristics of multi-physical quantity coupling in extreme environments constitute the core problem limiting measurement accuracy and stability. From the expression of total phase difference, nΔL represents the phase modulation due to cavity length change, and LΔn represents the phase modulation caused by refractive index change. Temperature, pressure, strain, vibration, material stress, and other factors all affect the F–P interference output by simultaneously modifying both L and n.

(1) Temperature cross-coupling mechanism

Temperature is the primary factor affecting the stability of fiber-optic F–P sensors. It induces cavity length changes via thermal expansion (as in Eq. (12)) and refractive index changes via the thermo-optic effect (as in Eq. (13)), thereby forming a dual modulation mechanism. The total optical path change is given by Eq. (14), and the corresponding wavelength shift by Eq. (15). Consequently, even during pressure or strain measurement, temperature variations introduce additional phase drift, resulting in temperature-induced cross-coupling. Under extreme high temperatures, parameters such as the coefficient of thermal expansion, refractive index, elastic modulus of packaging materials, and interfacial thermal stress exhibit nonlinear changes, leading to nonlinear drift of the interference spectrum and increased zero-point drift.

\Delta L=\alpha L\Delta T\;,\[\Delta L=\alpha L\Delta T\;,\] (12)

\Delta n=\xi n\Delta T\;,\[\Delta n=\xi n\Delta T\;,\] (13)

\Delta (OPD)=2(n\Delta L+L\Delta n) \;,\[\Delta (OPD)=2(n\Delta L+L\Delta n) \;,\] (14)

\Delta\lambda=(\alpha+\xi)\Delta T\; .\[\Delta\lambda=(\alpha+\xi)\Delta T\; .\] (15)

(2) Pressure and stress coupling mechanism

A diaphragm-type fiber-optic F–P pressure sensor relies on the flexural deformation of the diaphragm to change the cavity length. The cavity length change equals the central displacement of the diaphragm, satisfying Eq. (16) under small deformation. When temperature rises, Young's modulus decreases, the diaphragm stiffness decreases, and the pressure sensitivity changes accordingly, forming temperature-pressure coupling.

\delta \propto \frac{P{a}^{4}}{E{h}^{3}} \;,\[\delta \propto \frac{P{a}^{4}}{E{h}^{3}} \;,\] (16)

where P is the pressure; a is the diaphragm radius; h is the diaphragm thickness; E is Young's modulus.

(3) Vibration and dynamic phase disturbance mechanism

Vibration induces periodic changes in the cavity length (as in Eq. (17)), which correspond to dynamic phase modulation (as in Eq. (18)). Simultaneously, it may cause interference frequency shift, fiber bending loss, cavity surface tilt, and degradation of interference contrast.

L(t)={L}_{0}+\Delta L\sin ({\omega }_{v}t)\;,\[L(t)={L}_{0}+\Delta L\sin ({\omega }_{v}t)\;,\] (17)

\phi(t)=\frac{4\pi n}{\lambda}[L_0+\Delta L\sin(\omega_vt)]\; .\[\phi(t)=\frac{4\pi n}{\lambda}[L_0+\Delta L\sin(\omega_vt)]\; .\] (18)

Based on the above mechanisms, the interference phase change of fiber-optic F–P sensors in extreme environments results from the combined effects of multiple physical fields. The system output can be expressed as Eq. (19). Nonlinear coupling exists among the physical quantities. Traditional single-parameter calibration methods are insufficient to meet the demands of high-precision measurement. Therefore, it is necessary to suppress cross-interference through either multi-cavity structure decoupling or multi-physical field joint modeling.

\Delta\phi=\Delta\phi_T+\Delta\phi_P+\Delta\phi_{\varepsilon}+\Delta\phi_v+\Delta\phi_n\; .\[\Delta\phi=\Delta\phi_T+\Delta\phi_P+\Delta\phi_{\varepsilon}+\Delta\phi_v+\Delta\phi_n\; .\] (19)

The models described above capture the common coupling mechanisms through which multiple physical fields act on the F-P cavity. However, the actual degree of coupling is highly dependent on sensor material, cavity dimensions, sensing structure, and operating conditions, and thus cannot be characterized by a single unified coefficient. Consequently, multiparameter coupling should generally be evaluated on a case-by-case basis for specific application scenarios, using quantitative indicators such as cross-sensitivity and residual measurement error after decoupling. For instance, in combined temperature–pressure measurements, the pressure–temperature cross-sensitivity can be employed to quantify the equivalent pressure error induced by temperature variations. A value as low as 5.96 Pa/°C has been reported for an all-silicon dual-cavity F–P sensor⁶¹. For an F-P sensor that incorporates an independent temperature-compensation channel, the pressure measurement accuracy after compensation reached 0.52 %FS, providing another quantitative metric for assessing decoupling performance⁵⁶.

3 Chipization of sensitive structures, structurization of sensing chips, and metallization packaging of fiber-optic F–P sensors

3.1 Overview of measurable physical quantities and classification of fiber-optic F–P sensors

In recent years, the application scope of fiber-optic F–P sensors has expanded from traditional mechanical parameters to areas such as environmental sensing, field detection, and functional micro-devices. Driven by optical microcavities and high-sensitivity interferometric readout techniques, they have gradually become suitable for weak signal and complex scenario detection^23,24. Concurrently, the integration of F–P interference structures with specialty optical fibers, microcavities, and functional materials has promoted structural diversification in medium sensing and multi-physical quantity detection^25,26. Furthermore, the development of three-dimensional micro-structuring on fiber end faces and micro-packaging techniques has further advanced the evolution of sensors from single-cavity structures toward miniaturized, integrated, and functional devices^27,28.

Given the differences in action mechanisms, sensing principles, and application environments among various physical quantities, this paper categorizes representative works into two classes: mechanical parameter detection and environment/field interaction parameter detection^29–33. Their applications in structural deformation sensing, dynamic load measurement, functional material coupling, and complex environment monitoring are discussed separately. Combined with the related work of our research group, the research ideas toward multi-parameter synchronous measurement, chip-based design, and metallization packaging are outlined.

3.2 Fiber-optic F–P sensors for typical mechanical parameters

Detection of mechanical parameters represents an important application direction of fiber-optic F–P sensors, with related research mainly focusing on force, displacement, vibration, pressure, and strain. As shown in Fig. 3, owing to the distinct forms of action and structural response modes associated with different mechanical parameters, existing works have established multiple implementation paths in terms of sensitive structure design, cavity configuration, fabrication processes, and signal demodulation methods.

Figure 3. (a) 3D spring-based fiber-tip F–P micro-force sensor. (b) Cascaded F–P fiber strain sensor based on the comb-spectrum Vernier effect. (c) Sapphire-diaphragm fiber-optic F–P vibration sensor. (d) In situ 3D-printed fiber-tip F–P micro-interferometric accelerometer. (e) Cascaded sapphire-fiber F-P/FBG strain sensor for high-temperature strain measurement. (f) Temperature-compensated sapphire MEMS F–P pressure sensor. Fig. reproduced from: (a) ref.³⁴, John Wiley and Sons; (b) ref.⁴⁰, IEEE; (c) ref.³⁶, American Chemical Society; (d) ref.³⁷, Light Publishing Group; (e) ref.⁴¹, Elsevier; (f) ref.³⁹, Optica Publishing Group.

For micro-force measurement, Shang et al. (2024)³⁴ fabricated a three-dimensional spring-integrated F–P microcavity on the fiber end face using two-photon polymerization, and realized force detection based on cavity length change caused by spring compression. Experimental results show that the sensor has a force sensitivity of 0.436 ± 0.007 nm/nN, a resolution of 40.0 ± 0.7 pN, and a detection range of 34.8 nN.

For displacement measurement, Li et al. (2024)³⁵ proposed an ethanol-filled fiber microcavity displacement sensor. They utilized lateral 1550 nm laser-induced Marangoni effect to drive microbubble movement, and demodulation was achieved through optical path length change of the F-P cavity. The spatial frequency demodulation sensitivity in the range of 0–2550 μm is 0.00115 nm⁻¹/μm, the small-range wavelength demodulation sensitivity is 6.91 nm/μm, and the resolution is approximately 6 nm.

Vibration and acceleration detection typically require combining an inertial proof mass, diaphragm, or micro-beam structure with the F–P cavity to convert external dynamic loads into cavity length changes. Cao et al. (2024)³⁶ constructed a high-temperature fiber-optic F–P vibration sensor using a sapphire sensitive diaphragm. The sensor can operate at 600 °C, with a room temperature sensitivity of 38.66 nm/g, a characteristic frequency of 2446 Hz, and a transverse response ratio of 4.09%. Wang et al. (2025)³⁷ fabricated an F–P micro-interferometric acceleration sensor integrating a proof mass, reflective film, and supporting micro-beams on the end face of a ferrule by in-situ three-dimensional micro-printing. The sensor exhibits a linear response in the range of 0–10 g, a flat response bandwidth of 2–3 kHz, and a noise-equivalent acceleration of 62.45 μg/√Hz.

For pressure measurement, fiber-optic F–P sensors typically convert pressure into cavity length changes using elastic diaphragms, polymer spacer layers, or high-temperature resistant cavities. Zhen et al. (2023)³⁸ constructed a fiber end-face F-P pressure sensor using an atomically smooth gold micro-sheet and a PDMS spacer layer. The device with a PDMS clamped beam shows a sensitivity of 11.48 nm/kPa in the range of 0–2.34 kPa, with a pressure resolution of approximately 1.73 Pa. Liao et al. (2025)³⁹ fabricated a temperature-compensated fiber-optic F-P pressure sensor based on sapphire MEMS technology and direct bonding technology, capable of measuring temperature and pressure in the ranges of 25–1500 °C and 0–1 MPa. The pressure sensitivity is 0.782–0.952 μm/MPa, and the pressure measurement accuracy after compensation is 0.86% F.S.

For strain detection, cascaded F–P cavities, the Vernier effect, and temperature compensation structures are used to improve demodulation sensitivity or reduce thermal cross-sensitivity. Wei et al. (2024)⁴⁰ proposed a cascaded F–P cavity fiber-optic sensing structure based on the comb-spectrum Vernier effect to achieve strain and temperature demodulation. The strain and temperature sensitivities reach 4.76 pm/με and 271.61 pm/°C, respectively, which are approximately 21.6 times and 23.16 times higher than those of a single-stage cascaded F–P cavity. Shen et al. (2025)⁴¹ integrated a cascaded F–P cavity and a fiber Bragg grating into a sapphire fiber to achieve temperature-compensated strain measurement up to 1150 °C within a range of ±1000 με. The maximum measurement errors at room temperature and 1150 °C do not exceed 5% and 14%, respectively.

Overall, research on fiber-optic F–P sensors for typical mechanical parameters mainly focuses on the construction of micro-structured sensitive elements, the regulation of cavity response, and demodulation compensation under complex environmental conditions. The related work lay a foundation for subsequent sensor structural integration, multi-parameter measurement, and engineering packaging.

3.3 Fiber-optic F–P sensors for environment and field interaction parameters

With the development of sensitive materials, micro/nano fabrication, and end-face integration technologies, fiber-optic F–P sensors have expanded from typical mechanical parameter detection to environment and field interaction parameters, including acoustic waves, gases, magnetic fields, temperature, and humidity. As illustrated in Fig. 4, such sensors typically rely on acoustic-sensitive diaphragms, photothermal effects, magnetically sensitive units, thermally sensitive cavities, or hygroscopic materials for signal transduction. Their structural design increasingly depends on the compatibility between functional materials and microstructures. Among these, acoustic detection represents one of the key application directions of fiber-optic F–P sensors. Wang et al. (2025)⁴² fabricated a hollow F–P ultrasonic sensing structure on the fiber end face by two-photon polymerization and metal evaporation, and used it for photoacoustic imaging detection. The sensor exhibits an acoustic pressure sensitivity of 797 mV/kPa, a noise-equivalent pressure of 2.8 Pa, a center response frequency of 1.5 MHz, and a bandwidth of 1.2 MHz. Different from ultrasonic detection, Wang et al. (2024)⁸ proposed a miniature fiber-optic F–P acoustic sensor for mHz-level infrasound signals, which has a flat response in the range of 0.01–2500 Hz, an acoustic pressure sensitivity of −123.19 dB re 1 rad/μPa at 5 Hz, and a minimum detectable pressure of 1.2 mPa/√Hz.

Figure 4. (a) 3D-printed multicore fiber-tip F–P sensor for discriminative magnetic field and temperature measurements. (b) Fiber-tip 3D-microprinted F–P photothermal interferometric gas sensor. (c) Miniaturized fiber-optic F–P sensor for mHz infrasound detection. Figure reproduced from: (a) ref.⁴⁵, Light Publishing Group; (b) ref.⁴³, John Wiley and Sons; (c) ref.⁸, Optica Publishing Group/Chinese Laser Press.

Gas detection, in contrast, relies more heavily on photothermal effects, absorption spectra, and gas-cavity interaction within microcavities. Zhao et al. (2024)⁴³ fabricated a low-finesse F–P microcavity on the end face of a single-mode fiber by three-dimensional micro-printing and realized C₂H₂ detection based on the photothermal interference mechanism. The cavity length is approximately 66 μm, the noise-equivalent concentration is 160 ppb, the response time is less than 0.5 s, and the signal fluctuation over 25 h is approximately ±1.5%. Zhao et al. (2026)⁴⁴ introduced the mode Vernier effect to construct a short-cavity F–P photothermal detection unit, achieving C₂H₂ detection in a 1 mm cavity length and 0.8 nL hollow-core gas cell. The lowest detection limit is 12 ppb, the noise-equivalent concentration is 75 ppb, and the response time is approximately 3 s.

Magnetic field and temperature detection are often challenged by temperature-induced cross-sensitivity and the need for multi-parameter decoupling. Consequently, recent studies have employed multi-channel structures on multi-core fiber end faces to achieve discriminative measurement. Xiong et al. (2024)⁴⁵ fabricated dual F–P microcavities on the end face of a multi-core fiber using two-photon polymerization to achieve magnetic field and temperature measurement. The maximum magnetic field sensitivity is 1805.6 pm/mT, the response time is approximately 213 ms, and the temperature sensitivity is 160.3 pm/°C. For simultaneous temperature and pressure measurement, Chen et al. (2024)⁴⁶ fabricated an open F–P microcavity by femtosecond laser two-photon 3D printing, and compensated for temperature crosstalk using an in-fiber FBG to achieve simultaneous measurement of temperature and pressure. The pressure sensitivity is 6.6649 nm/MPa, and the temperature sensitivities of the F-P microcavity and FBG are 0.105 nm/°C and 9.5 pm/°C, respectively.

For humidity detection, the adsorption, swelling, and mass loading changes of the sensitive film directly affect the sensor output. Li et al. (2026)⁴⁷ constructed a Ti₃C₂Tₓ MXene-GO composite film fiber-optic F–P resonant humidity sensor, achieving humidity detection through photothermal excitation and non-contact interferometric readout. In the range of 2–100%RH, the absolute sensitivity and relative sensitivity are 2.1 kHz/%RH and 0.22%/%RH, respectively, and the response/recovery time is 3.61/5.53 s.

These findings indicate that the development of fiber-optic F–P sensors for environmental and field interaction parameters not only expands the range of measurands but also drives structural evolution toward configurations based on sensitive diaphragms, photothermal microcavities, functional materials, and resonant units. Compared with sensors designed for typical mechanical parameters, such devices are more susceptible to material response, environmental disturbances, and cross-sensitivity, requiring further improvements in structural stability, decoupling capability, and packaging adaptability.

3.4 Fiber-optic F–P sensors for multi-parameter synchronous testing

In extreme environments such as nuclear energy facilities, aero-engines, and high-temperature fluid equipment, measurement of a single physical quantity often fails to meet the requirements of practical condition monitoring¹². Fiber-optic F–P sensors offer advantages such as small size, immunity to electromagnetic interference, long-distance transmission capability, and ease of integration with high-temperature resistant materials and metal packaging structures, rendering them well suitable in high-temperature, high-pressure, highly corrosive, and confined installation spaces^48–51. Temperature, pressure, strain, displacement, and vibration often act simultaneously on the sensor structure and introduce strong thermal drift, packaging stress, and cross-sensitivity. Consequently, compared with the aforementioned single-parameter fiber-optic F–P sensors, engineering applications in extreme environments place greater emphasis on long-term stability of the sensor under complex service conditions, while requiring in-situ temperature information during the measurement of the target parameter, and achieving temperature compensation through structural design and demodulation algorithms. As shown in Fig. 5, to address this issue, our group has carried out a series of studies ranging from single-parameter adaptation to extreme environments to composite-cavity multi-parameter synchronous measurement. This work has led to a design approach for fiber-optic F–P sensors characterized mainly by chipization of sensitive structures, structurization of sensing chips, and all-metal packaging.

Figure 5. (a) Miniaturized fiber-optic F–P force sensor for radial impact force measurement. (b) Miniaturized fiber-optic F–P displacement sensor for fuel-pin vibration monitoring. (c) All-metal packaged temperature-compensated fiber-optic F–P strain sensor for high-temperature liquid-metal environments. (d) Dual-function composite-cavity fiber-optic F–P sensor for simultaneous pressure and temperature measurement. (e) All-sapphire composite-cavity fiber-optic F–P pressure sensor system with in situ temperature compensation. (f) Composite-cavity fiber-optic F–P interferometric accelerometer with temperature calibration for high-temperature and high-pressure applications. Figure reproduced from: (a) ref.⁵², MDPI; (b) ref.⁵³, John Wiley and Sons; (c) ref.⁵⁵, Elsevier; (d) ref.⁵⁶, IEEE; (e) ref.⁵⁷, Springer Nature; (f) ref.⁵⁸, Springer Nature.

Early work mainly focused on the measurement of mechanical parameters in nuclear power equipment under confined spaces, high temperature, high pressure, and humid/hot environments. We (2017)⁵² designed a miniaturized fiber-optic F–P force sensor for detecting radial collision force between steam generator heat transfer tubes and support plates. The packaged sensor dimensions were 17 mm × 5 mm × 3 mm, and it was tested under humid/hot conditions, high pressure of 10 MPa, high temperature of 350 °C, and vibration environment of 40 kHz. Subsequently, the sensor was installed in a 1:1 steam generator test loop to acquire radial collision force signals. This work focused more on the design of force sensing structures in confined spaces and high-temperature high-pressure packaging adaptation, laying a foundation for subsequent engineering applications of fiber-optic F–P devices in nuclear power scenarios.

For monitoring fuel rod vibration displacement, we (2024)⁵³ designed a cylindrically packaged fiber-optic F–P displacement sensor with an overall device size of 30 mm × 6.5 mm. The F–P cavity was formed jointly by a sapphire glass sheet and the surface of the fuel rod under test. Experimental results show that the displacement measurement range of the sensor in a wet environment is approximately 1250 μm, and the RMSE in wet environment and at 350 °C high temperature are about 0.112 and 0.144, respectively, indicating that such sensors have begun to shift from simple structural feasibility verification to measurement stability verification under real service environments.

For strain detection of fuel assemblies, we (2022)⁵⁴ proposed a fiber-optic F–P strain sensing system based on non-scanning correlation demodulation and designed a dual-elastic-ring structure to transfer the strain on the fuel plate surface to the F–P cavity length change. Experimental results show that the strain sensitivity of the system at 300 °C reaches 12.6 nm/με, with a dynamic test frequency range of 10–500 Hz. In thermal-hydraulic experiments under high temperature, high pressure, and high-speed water flow scouring, the consistency deviation of the measurement data was less than 1.5%.

Building on this foundation, the research group further shifted its focus to composite cavity structures and temperature compensation methods. We (2025)⁵⁵ proposed an all-metal packaged temperature-compensated fiber-optic F–P strain sensor for surface strain monitoring in a liquid metal environment at 500 °C and 2 MPa. A composite cavity structure, consisting of a metal hollow strain tube, a metal ferrule, and a sapphire substrate was employed to achieve strain measurement and in-situ temperature monitoring. The sensor uses 316L stainless steel as the packaging material, combined with nano-silver low-temperature sintering and laser welding processes to achieve high-temperature sealing. A strain sensitivity of 2.43 nm/με was obtained in a liquid metal environment at 500 °C, and the cavity length vibration remained below 8 nm over a 60-hour high-temperature and high-pressure stability test.

Simultaneous measurement of pressure and temperature is a typical application direction of composite-cavity fiber-optic F–P sensors. We (2025)⁵⁶ designed an integrated temperature-pressure composite-cavity fiber-optic F–P sensor, in which a substrate cavity and an air cavity carry temperature and pressure information respectively, and fast dual-cavity decoupling was achieved through dedicated spectrometer hardware design and spectral region separation. In experiments, the sensor achieved simultaneous temperature and pressure measurement in the ranges of 0–500 °C and 0–4 MPa, with a temperature measurement accuracy of 0.35%FS and a pressure measurement accuracy after temperature compensation of 0.52%FS.

For pressure measurement under even more extreme temperature conditions, we (2026)⁵⁷ further proposed an all-sapphire composite-cavity fiber-optic F–P high-temperature pressure sensing system. The sensing chip adopts a dual-cavity structure consisting of a sapphire substrate cavity and an air pressure cavity, and a pressure-sensitive diaphragm with a central platform is used to improve the reflection spectrum quality. In terms of fabrication, the system combines MEMS wet etching, RIE hard mask preparation, and high-temperature wafer-level bonding to achieve sapphire sensitive chip processing. Meanwhile, the APSC-FFT algorithm is proposed for cavity length demodulation. Experimental results show that the system can operate in the ranges of 28–800 °C and 0–1.2 MPa, with temperature and pressure system errors better than 0.13%FS and 0.18%FS, respectively, and pressure stability better than 0.12%FS. The unpackaged chip retains a measurable signal after long-term annealing at 1500 °C.

In addition to static pressure and strain measurement, composite cavity structures have also been employed for dynamic acceleration sensing. We (2026)⁵⁸ proposed a composite-cavity fiber-optic F–P acceleration sensor with temperature calibration function for monitoring flow-induced vibration of steam generator heat transfer tubes in pressurized water reactors. In this sensor, a glass substrate cavity is used for temperature measurement, and an air cavity is used for acceleration response. The sensing chip adopts a three-layer structure consisting of a silicon diaphragm and a glass substrate, and chip fabrication and high-temperature sealing are achieved through wet etching, silicon-glass bonding, and laser welding. Experimental results show that the sensor exhibits a room temperature sensitivity of 4.53 nm/g, a resonant frequency of 7450 Hz, a minimum transverse sensitivity of 0.281%, and a cavity length drift of less than 0.1 nm during a 60-hour continuous test at 350 °C and 17.5 MPa.

The foregoing work demonstrates that the research group’s contributions extend beyond the mere deployment of fiber-optic F–P sensors for the detection of diverse physical quantities. Instead, they systematically address common challenges inherent to extreme-environment measurements, including temperature crosstalk, cavity length drift, structural reliability, and packaging adaptability. Early force and displacement sensors mainly solved the problems of structural miniaturization and engineering installation under high-temperature, high-pressure, humid/hot environments. Subsequent strain, pressure, and acceleration sensors further introduced composite cavity structures, employing an air cavity for target parameter measurement and a substrate cavity for in-situ temperature characterization. Through spectral separation, FFT demodulation, temperature calibration models, and all-metal packaging, these sensors have achieved improved measurement stability in complex environments. Overall, this line of work reflects the technical path of fiber-optic F–P sensors evolving from single sensitive cavities to composite cavities, multi-parameter synchronous measurement, and extreme-environment packaged devices, and also provides a foundation in structural design and fabrication processes for subsequent multi-physics collaborative sensing in high-temperature high-pressure environments.

To further illustrate the application characteristics of composite cavity structures in multi-parameter measurement in extreme environments, this paper selects representative F-P sensors for high-temperature strain, pressure/temperature synchronous measurement, and acceleration detection for comparison^59–63, as shown in Table 1. Since the sensitivity units and evaluation indicators corresponding to different measurement objects are not consistent, simply comparing sensitivity magnitudes cannot fully reflect sensor performance. Therefore, the table mainly summarizes the working conditions, key performance, temperature compensation methods, and packaging adaptability of the sensors.

Table 1. Comparison of representative F–P sensors for multiparameter sensing in extreme environments.

Although composite cavities offer an effective means for multiparameter simultaneous measurement using F-P sensors, several common challenges persist, including cross-sensitivity, long-term stability, and engineering practicality. Different measurands influence the interference signal through variations in cavity length, refractive index, and sensing-structure characteristics. Moreover, similar optical path differences or overlapping spectral features among multiple cavities further complicate parameter separation and stable demodulation. Harsh environments involving high temperature, high pressure, and corrosion can alter material properties and packaging stress, leading to sensitivity drift. In addition, the trade-off between high-accuracy demodulation and high-speed dynamic response, along with issues in miniaturized packaging, manufacturing consistency, and repeatable calibration, remains a significant hurdle for practical deployment.

Looking ahead, multiparameter F-P sensing may advance through composite cavities or hybrid structures with differentiated responses, combined with intelligent algorithms and physical models to enhance parameter decoupling under complex operating conditions. Materials resistant to extreme environments, such as sapphire and ceramics, together with advanced microfabrication techniques, can further improve environmental adaptability and structural integration. Meanwhile, on-chip spectral analysis and photonic integration are expected to drive the miniaturization and low-power development of interrogation systems, while also facilitating integrated sensing of temperature, pressure, strain, and dynamic parameters.

4 Key technologies of demodulation systems for fiber-optic F–P sensors

4.1 Technical routes for fiber-optic F–P demodulation

Fiber-optic F–P sensors feature a simple structural configuration. The sensing unit generally consists of a F–P cavity formed by two or more reflective surfaces. Figure 6 shows a typical reflection-type structure. Incident light enters the cavity and is reflected by multiple reflective surfaces to form an interference signal. Under low finesse conditions, when only two-beam interference is considered, given the incident light intensity I_in, wavelength λ, reflectivity of the first surface R0, reflectivity of the second surface R₁, F–P cavity length L₁, refractive index n₁, the reflected light intensity model for a single cavity I_{single_cavity}, is expressed as follows:

Figure 6. F–P sensor basic demodulation technology roadmap.

\begin{split}I_{\mathrm{single}_{\mathrm{cavity}}}= & R_0I_{\mathrm{in}}+(1-R_0)^2R_1I_{\mathrm{in}} \\ & +2\sqrt{R_0R_1}(1-R_0)\mathrm{cos}(\frac{4\pi n_1L_1}{\lambda}))I_{\mathrm{in}}\;.\end{split}\[\begin{split}I_{\mathrm{single}_{\mathrm{cavity}}}= & R_0I_{\mathrm{in}}+(1-R_0)^2R_1I_{\mathrm{in}} \\ & +2\sqrt{R_0R_1}(1-R_0)\mathrm{cos}(\frac{4\pi n_1L_1}{\lambda}))I_{\mathrm{in}}\;.\end{split}\] (20)

Based on the reflected light intensity model, the development of cavity length calculation schemes can be roughly divided into intensity demodulation, phase demodulation, spectral demodulation, and other schemes developed in recent years. Early traditional intensity demodulation schemes used monochromatic light as the light source, utilizing the variation of interference light intensity with optical path difference. Cavity length demodulation could be achieved by detecting the light intensity. The demodulation speed is fast, but the single dynamic range is small, and it is easily affected by the environment^64–66. The phase demodulation algorithm for monochromatic light offers higher accuracy, but may suffer from 2π ambiguity and is easily affected by the environment^67–68. Around 2000, with the development of broadband light sources, spectrometers, detectors, etc., spectral-based demodulation schemes began to develop, including spectral-domain peak tracking, Fourier transform, spectral comparison, etc.^69–70. Among them, white light interferometry offers improved stability, and the spectral comparison approach, in particular, provides extremely high accuracy, making it the current mainstream high-precision calculation scheme. However, spectral comparison is computationally intensive and relatively slow.

Current research on demodulation schemes for F–P-type sensing systems is mainly developing in the direction from high sensitivity to high accuracy, and then to high accuracy, multi-parameter, and intelligent approaches⁷¹. Relying on the development of MEMS technology, the fabrication of F–P-type sensors has become more convenient, leading to the emergence of F–P sensors with composite cavity structures for simultaneous measurement of temperature-pressure, temperature-salinity, temperature-refractive index, etc.^61,72,73. The composite cavity model is as follows:

\begin{split} & I_{\mathrm{compound}_{\mathrm{cavity}}}(\lambda)=R_0I_{\mathrm{in}}+(1-R_0)^2R_1I_{\mathrm{in}}+ \\ & (1-R_0)^2(1-R_1)^2R_2I_{\mathrm{in}} \\ &\left. +2\sqrt{R_0R_1}(1-R_0)\mathrm{cos}\left(\frac{4\pi n_1(\lambda)L_1}{\lambda}\right)\right)I_{\mathrm{in}} \\ & +2\sqrt{R_1R_2}(1-R_0)^2(1-R_1)\mathrm{cos}\left(\frac{4\pi n_0L_2}{\lambda}\right)I_{\mathrm{in}} \\ & +2\sqrt{R_0R_2}(1-R_0)(1-R_1)\mathrm{cos}\left(\frac{4\pi n_1(\lambda)L_1}{\lambda}+\frac{4\pi n_0L_2}{\lambda}\right)I_{\mathrm{in}}\ .\end{split}\[\begin{split} & I_{\mathrm{compound}_{\mathrm{cavity}}}(\lambda)=R_0I_{\mathrm{in}}+(1-R_0)^2R_1I_{\mathrm{in}}+ \\ & (1-R_0)^2(1-R_1)^2R_2I_{\mathrm{in}} \\ &\left. +2\sqrt{R_0R_1}(1-R_0)\mathrm{cos}\left(\frac{4\pi n_1(\lambda)L_1}{\lambda}\right)\right)I_{\mathrm{in}} \\ & +2\sqrt{R_1R_2}(1-R_0)^2(1-R_1)\mathrm{cos}\left(\frac{4\pi n_0L_2}{\lambda}\right)I_{\mathrm{in}} \\ & +2\sqrt{R_0R_2}(1-R_0)(1-R_1)\mathrm{cos}\left(\frac{4\pi n_1(\lambda)L_1}{\lambda}+\frac{4\pi n_0L_2}{\lambda}\right)I_{\mathrm{in}}\ .\end{split}\] (21)

For the calculation of composite cavities, decoupling of the two-cavity information is one of the most important steps. Through Fourier transform, time-frequency analysis, etc., separate extraction and decoupling of the two cavities can be achieved⁵⁶. In recent research on demodulation schemes, neural networks have begun to be directly used as demodulation tools. Deep learning-based methods allow a network to directly learn the mapping from spectra to cavity length (or multiple cavity lengths) from a large amount of experimental spectral data^74,75.

4.2 Demodulation techniques for single-cavity fiber-optic F–P sensors

4.2.1 Intensity demodulation techniques

Among the demodulation techniques for F–P-type sensors, intensity demodulation offers advantages such as system simplicity, fast response speed, and low cost⁷⁶. However, issues such as the cosine nonlinearity between interference light intensity and cavity length, the sensitivity of the static operating point to ambient temperature, and fluctuations in optical path loss have long constrained the accuracy and stability of traditional intensity demodulation methods. The core of traditional intensity demodulation is to operate the system at the "quadrature point" (Q-point) of the interference curve, as shown in Fig. 7(a), where the slope is maximum (approximately linear region). When an external signal causes a cavity length change, the reflected light intensity is approximately linearly related to the cavity length (ΔL = αΔI), where α is the detection sensitivity⁶⁴. To overcome the aforementioned limitations, researchers have developed various improved techniques, including operating point control methods, phase quadrature stabilization methods, multi-wavelength intensity demodulation, and harmonic component amplitude ratio methods.

Figure 7. (a) Relationship between the F–P cavity length and the interference light intensity, (b) Harmonic amplitude ratio demodulation system and Curve of R_mn varying with D, (c) phase-shifting optical demodulation system, (d) acoustic signal testing system based on four-wavelength demodulation. (e) Non-scanning optical correlation demodulation system, (f) dual Fizeau demodulation system. Figure reproduced from: (a) ref⁷⁶, IEEE; (b) ref⁶⁴, Optica Publishing Group; (c) ref⁸², IEEE; (d) ref⁸³, Optica Publishing Group; (e) ref⁸⁹, Optica Publishing Group; (f) ref⁹⁰, Springer Nature.

To break the strict dependence of the operating point control method on the Q-point and expand the demodulation range, in 2024, Qiu et al. proposed a demodulation algorithm based on the amplitude ratio of harmonic components (ARHC). This method uses Bessel function expansion of the interference signal and obtains the intensity ratio of the fundamental frequency to the double frequency via Fourier transform. This ratio is related only to the amplitude of the signal under test and is independent of the initial phase introduced by the static cavity length. Figure 7(b) shows the acquisition system and the relationship among the amplitude ratio, cavity length, and harmonic components. This method achieves a larger demodulation range, higher accuracy, and better environmental adaptability.

Zuo et al.(2024) proposed a magnetic field sensor based on dual-wavelength elliptical demodulation. An external magnetic field changes the complex refractive index of the magnetic fluid, thereby modulating the reflected light intensities at two wavelengths. The two light intensities satisfy an elliptical function relationship, and the semi-major axis of the ellipse is inversely proportional to the magnetic field strength. This scheme exhibits good linearity in the range of 0–50 Gs, with a measurement accuracy of 2.62 Gs, and the ratio processing effectively suppresses light source fluctuations and optical path disturbances⁷⁷.

Li et al. (2025) applied feedback-type single-wavelength intensity demodulation technology to real-time gear fault detection, the reflected light is received by a photodetector, and an FPGA analyzes the demodulated signal in real time and controls the filter voltage to dynamically lock the Q-point. The system acquires vibration signals of 27 gear health states at a sampling rate of 100 kHz⁷⁸.

To overcome the problems of light source fluctuation, fiber disturbance, and Q-point drift caused by temperature in traditional intensity demodulation, Ye et al. (2026) proposed a high-stability quadrature point demodulation technique. Common-mode interference is eliminated by calculating the ratio of the sensing signal to the reference signal, while the low-frequency zero-drift component is extracted to provide feedback control of the filter center wavelength, actively compensating for Q-point drift. The nonlinear error is less than 1.5% in the range of 25 °C to 300 °C⁷⁹.

For multi-channel scenarios, Wang et al. (2026) designed an intensity demodulation system integrating a biconical MZI sensor and a ceramic ferrule FPI sensor. The FPI sensor utilizes the characteristic that its light intensity varies periodically with temperature to achieve high-precision measurement in critical intervals such as 44 °C to 46 °C (detection limit 0.02 °C)⁸⁰.

Intensity demodulation offers a simple configuration and fast response, making it well suited for rapid dynamic measurements where cavity-length variations remain small and within an approximately linear range. In addition, maintaining the operating point near the quadrature point enables high sensitivity.

4.2.2 Phase demodulation

Unlike intensity demodulation, phase demodulation directly extracts the phase change of the interference light, fundamentally avoiding light intensity fluctuations and nonlinear distortion. According to the carrier generation method, it can be divided into two major categories: active carrier modulation and passive phase shift detection. The phase-generated carrier (PGC) is an active modulation method that shifts the signal spectrum to the high-frequency region by introducing a high-frequency carrier in the reference arm, thereby suppressing low-frequency environmental noise. An improved Arctan demodulation scheme proposed by Zhang et al. (2026). Orthogonal signals independent of θ and C were constructed using the sine and cosine components of the fundamental frequency and double frequency and their differentials, with a measured total harmonic distortion below 0.5%⁸¹.

As shown in Fig. 7(c)⁸², Liu et al.(2021) used a Faraday Michelson interferometer with passive demodulation based on a 3×3 coupler, requiring no external modulation. The 3×3 coupler and two Faraday mirrors formed a demodulator with a fixed phase shift of 120°. Combined with linear fitting and the DCM algorithm, they successfully demodulated a large vibration signal of 3.5 kHz with an amplitude of 12.9 μm. Shi et al. (2022) proposed a single-wavelength passive phase shift demodulation technique. After eliminating the DC and AC amplitude differences using ellipse fitting, the phase was demodulated by arctangent. This scheme achieved linearity R² = 0.99993 in the amplitude range of 0.01–2250 nm⁸⁴.

Huang et al.(2024) proposed an improved dual-wavelength method based on prior information. A reference strong ultrasound signal was introduced to pre-construct a complete Lissajous figure, and ellipse fitting was used to obtain parameters of DC, AC, and phase difference. Linearity reached 0.9989 in the sound pressure range of 0.20–13.36 Pa⁸⁵. Zhang et al. (2024) proposed a quadrature-phase three-wavelength demodulation method. The three interference signals differ sequentially by (2m＋1/2)π and (2m＋1)π. After eliminating the DC component by linear fitting, the phase was demodulated using arctangent⁸⁶. Zhang et al.(2022) proposed a four-wavelength quadrature phase demodulation technique suitable for extrinsic F–P interferometer (EFPI) sensors and dynamic signals, as shown in Fig. 7(d). Four wavelengths were selected according to the free spectral range (FSR) of the EFPI. Linear fitting was used to obtain the linear equations of two sets of anti-phase signals, which were normalized to obtain standard orthogonal signals, and then the phase was demodulated by differential cross-multiplication. Experiments showed linearity of 0.9999 in the sound pressure range of 0.24–6.19 Pa at 250 Hz, with a frequency response covering 1 Hz to 100 kHz⁸³.

Phase demodulation is well suited for displacement and vibration measurements that demand high resolution, linearity, and fast dynamic response, while quadrature-phase detection effectively mitigates the influence of optical-intensity fluctuations. However, because monochromatic phase measurements suffer from a 2π ambiguity, determining the absolute cavity length requires either phase unwrapping or the use of additional absolute-reference information.

Monochromatic light phase demodulation utilizes the high coherence of lasers to achieve high-precision measurement via phase extraction. However, narrow-linewidth light sources cannot provide absolute cavity length information. With the development of broadband light sources, the cross-correlation demodulation method for white light interferometry has been developed. The basic principle involves using a reference interferometer. When the reference optical path length exactly matches the sensing optical path length, the position of the zero-order correlation peak is the absolute cavity length, free from periodic ambiguity constraints. Correlation demodulation is primarily classified into scanning and non-scanning types. Scanning correlation demodulation changes the reference optical path length through mechanical or light source tuning in the time dimension to achieve matching⁸⁷. In contrast, non-scanning correlation demodulation expands different optical path differences spatially at once, with different positions on the detection array corresponding to different reference cavity lengths⁸⁸.

We (2022) designed a non-scanning hardware optical correlation system based on a Fizeau interferometer for the dynamic detection of mask-wafer gap in surface plasmon lithography, as shown in Fig. 7(e). The light intensity corresponding to each pixel on the CCD is the optical cross-correlation result between the sensing cavity length and the two beams of light corresponding to the wedge thickness. The demodulation system achieved a resolution better than 0.30 nm and a static stability of ±0.12 nm in the range of 3.0–6.0 μm⁸⁹. We (2023) proposed a dual Fizeau interferometer demodulation system, as shown in Fig. 7(f). Two orthogonal signals can be used to calculate the phase, and then the absolute cavity length is obtained from the periodic relationship between the dual-wavelength phase difference and the cavity length, extending the limited range of single-wavelength demodulation to 4.74 μm, achieving large-range absolute measurement⁹⁰.

4.2.3 Spectral demodulation

In addition to intensity demodulation and phase demodulation, another major class of methods directly processes the reflection spectrum of the sensor in the wavelength domain. These methods calculate the cavity length by extracting characteristic features in the spectrum, such as extreme wavelength, frequency components, or correlation peaks. Typically, such methods employ broadband light sources or tunable lasers, combined with spectrometers or scanning filters to obtain interference spectra, thereby enabling absolute cavity length measurement that is immune to fluctuations in light intensity. However, the demodulation speed is limited by the spectral acquisition rate. These methods can be roughly divided into spectral peak tracking, Fourier transform, spectral comparison (cross-correlation), etc.⁹¹. The most direct spectral demodulation method is to track the spacing between adjacent peaks or dips in the interference spectrum. The cavity length d and the wavelength spacing Δλ between adjacent peaks (or dips) satisfy d = λ2/(2nΔλ). Jiang (2008) used a tunable F–P filter (FFP-TF) to scan the spectrum and calibrated the wavelength with an acetylene gas cell, improving the measurement resolution from 25 μm to 1 μm⁹². Chen et al. (2020) proposed a square peak-to-peak algorithm. The reflection spectrum was transformed to the frequency domain, the DC component was filtered out, and then squared. For sensors with a cavity length of 15–25 μm, the measured maximum error was only 0.030 μm⁹¹. Mei et al. (2015) proposed a multi-extremum tracking technique. A tunable laser was used to scan 1525–1565 nm with high precision to obtain the exact wavelengths of all 33 peaks and 32 dips. The interference order was obtained by linear fitting of all extrema, achieving a resolution of 1 nm and a dynamic range covering 250–3000 μm⁹³.

The Fourier transform method converts the wavelength-domain interference spectrum into the frequency domain via FFT, where the position of the main peak corresponds to the cavity length. This method offers fast processing speed, but the resolution is limited by the spectral sampling range and number of points. As shown in Fig. 8(a), Hai et al. (2024) proposed an improved algorithm combining FFT with mean square error (MSE) for a sapphire sandwich-structure F-P pressure sensor, combining the Fourier transform method and spectral comparison. The fitting error under a pressure of 0–10 MPa was less than 0.025%⁹⁴. The spectral comparison method performs a correlation operation between the measured spectrum and a series of ideal spectra with known cavity lengths. The cavity length corresponding to the correlation peak is the measured value. This method inherently offers noise immunity and sub-pixel accuracy. Shao et al. (2021) used cross-correlation to demodulate a sapphire composite cavity, obtaining a pressure sensitivity of 355.8 nm/MPa and a temperature sensitivity of 1.64 nm/°C, as shown in Fig. 8(b)⁹⁵. Li et al. (2024) used a 48-channel arrayed waveguide grating (AWG) as a spectroscopic element to perform FFT, frequency estimation, and phase calculation in real time, achieving a demodulation rate of 200 kHz, a cavity length range of 60–700 μm, and a resolution better than 0.22 nm⁹⁶.

Figure 8. (a) Testing system of sapphire sandwich-structured F–P pressure sensor, (b) high-temperature pressure sensing system and cross-correlation function of two ranges; (c) Beam propagations at a Fabry–Perot etalon and its fringe spectrum, (d) Intensity Demodulation of Optical Vernier Effect; (e) Single-parameter pre-training branch model. Figure reproduced from: (a) ref⁹⁴, MDPI; (b) ref⁹⁵, Optica Publishing Group; (c) ref⁹⁷, Optica Publishing Group; (d) ref⁹⁸, IEEE Xplore; (e) ref¹⁰⁰, Elsevier Ltd.

Spectral demodulation extracts the absolute cavity length from characteristic wavelengths, spatial-frequency components, or the full profile of a broadband interference spectrum. Its performance is governed primarily by the spectral bandwidth, resolution, sampling interval, and fringe visibility. Among the commonly used techniques, peak/valley tracking works well for static or slowly varying spectra that exhibit well-defined extrema and high visibility; Fourier-transform-based methods are effective for broadband, approximately periodic spectra and can resolve composite cavities when their frequency-domain peaks are distinguishable; and spectral comparison/cross-correlation approaches leverage the entire spectral shape, offering higher accuracy for absolute cavity-length measurements at the expense of increased computational cost.

4.2.4 Other demodulation schemes

In recent years, several improved techniques incorporating new principles or new tools have emerged. They either improve accuracy based on traditional methods or leverage optoelectronic devices and intelligent algorithms to expand system capabilities, providing more options for engineering applications of fiber-optic F–P sensors. Matsukuma et al. (2023) proposed an improved scheme spanning multiple free spectral ranges, as shown in Fig. 8(c). Two peaks separated by an integer k FSRs were selected. Using an air-gap F-P etalon with FSR ≈ 200 GHz and N = 62, after optimizing with k = 31, the standard deviation of repeated measurements was only 3.8 nm⁹⁷. Zhang et al. (2024) combined the optical Vernier effect with a three-layer backpropagation neural network, as shown in Fig. 8(d). Using only the relative light intensity ratios at three wavelengths, the interferometer's optical path difference was output after training, with a relative error ≤ ±1.23%⁹⁸. Chen et al. (2025) used an arrayed waveguide grating (AWG) and a convolutional bidirectional gated recurrent unit (CNN BiGRU) network, converting wavelength shifts into 28-channel transmitted light intensities. The network directly predicted three characteristic wavelengths from the multi-channel light intensity values⁷⁴. Li et al. (2026) arranged six micro-sphere FPI pressure sensors on the seat surface and backrest key pressure areas, using a MEMS optical switch for time-division multiplexing⁹⁹.

In addition to state recognition, deep learning can also automatically extract parameter features directly from raw spectra, making it particularly suitable for multi-parameter measurement where cross-sensitivity exists. As shown in Fig. 8(e), Li et al. (2026) constructed a dual-branch DenseNet network for a cascaded EFPI-FBG sensor. The final mean absolute error for temperature was 0.907 °C, and for pressure it was 0.035 MPa, with a single-spectrum demodulation time of 0.38 ms¹⁰⁰.

To achieve multi-parameter measurement or in-situ temperature compensation for F–P sensors, composite cavity sensors have been developed, integrating multiple parameters such as temperature and pressure into a single sensing head. Wang et al. (2021) directly bonded a silicon-based pressure-sensitive diaphragm to a silicon substrate to form a dual-cavity temperature-pressure sensor. The demodulation system and spectral schematic are shown in Fig. 9(a). A bandpass filter was used to separate the independent spectrum of each sub-cavity, and then the peak tracking method was used to accurately extract the cavity length. The pressure-temperature cross-sensitivity of this sensor was as low as 5.96 Pa/°C⁶¹.

Figure 9. Composite F–P cavity sensors. (a) All-silicon temperature–pressure composite cavity and Fourier transform demodulation, (b) sapphire–metal sheet temperature–pressure composite cavity and time–frequency analysis decoupling, (c) all-sapphire temperature–pressure composite cavity and peak-shift-corrected Fourier transform, (d) silica temperature–acceleration composite cavity and Fourier transform demodulation. Figure reproduced from: (a) ref⁶¹, Optica Publishing Group; (b) ref⁵⁶, IEEE; (c) ref⁵⁷, Springer Nature; (d) ref⁵⁸, Springer Nature.

In 2025, our research group also designed a sapphire-metal composite cavity sensor for composite-cavity temperature-pressure sensing. The sensor consists of a sapphire substrate and a metal diaphragm⁵⁶. As shown in Fig. 9(b), by controlling the resolution of the spectrometer, the information from the two cavities was segmented along the wavelength on the spectrometer. After confirming the segmentation positions through time-frequency analysis, the interference signals of the two sub-cavities were decoupled. In the ranges of 0–500 °C and 0–4 MPa, the temperature accuracy was 0.35% FS, and the pressure accuracy was 0.52% FS. To further improve high-temperature performance, in 2026 we developed an all-sapphire composite cavity sensor and proposed an adaptive peak-shape-corrected FFT algorithm (APSC-FFT)⁵⁷. As shown in Fig. 9(c), by iteratively adjusting the resampling interval in the wavenumber domain to make the FFT peak symmetric, the demodulation accuracy of the Fourier transform decoupling algorithm was further improved. In the ranges of 28–800 °C and 0–1.2 MPa, the temperature system error was 0.13% FS, and the pressure error was 0.18% FS. In the same year, we extended this composite cavity design concept to acceleration measurement, as shown in Fig. 9(d), adopting a three-layer chip structure: a glass substrate forms the temperature cavity, and an air cavity is formed between the glass and the proof mass⁵⁸. For demodulation, FFT combined with bandpass filtering was employed to separate the spectra, the peak tracking method was used to extract the cavity length, and real-time temperature compensation was performed through calibration curves. In a 60-hour stability test at 350 °C and 17.5 MPa, the cavity length drift was less than 0.1 nm, and the cross-axis sensitivity was 0.281%.

Machine learning offers a direct approach to establishing nonlinear mappings between F–P spectra and cavity length or other target physical quantities. However, its practical deployment demands careful consideration of training-data representativeness, model generalization, and real-time performance. To minimize the model's dependence on specific experimental conditions, the training data should encompass a wide range of measurement ranges, signal-to-noise ratios, and operating environments (e.g., temperature and pressure). Moreover, sensor-to-sensor fabrication variations and long-term drift in the interrogation system can cause distribution shifts between training and actual measurement data, underscoring the need to further enhance model robustness across different sensors and operating conditions^130,131. For dynamic and embedded demodulation applications, it is also essential to balance demodulation accuracy with real-time performance by carefully evaluating model size, computational latency, and hardware resource consumption.

4.3 Analysis of demodulation techniques in special environments

The application of fiber-optic F–P sensors in extreme environments characterized by high temperature, high pressure, strong electromagnetic interference, and confined spaces imposes stringent requirements on the stability, accuracy, dynamic range, and real-time performance of demodulation methods. Different demodulation techniques differ in their principles, system complexity, environmental adaptability, and performance metrics, and thus must be selected according to specific operating conditions. Intensity demodulation directly reflects cavity length changes by detecting variations in interference light intensity. It features the simplest system structure and the fastest response speed, but is susceptible to light source fluctuations and environmental disturbances. Phase demodulation extracts the phase change of the interference signal, offering good linearity and a large dynamic range, but with relatively higher system complexity. Spectral-domain demodulation directly uses the reflection spectral characteristics of a broadband light source or tunable laser to calculate the absolute cavity length, without being affected by light intensity fluctuations, but the demodulation speed is limited by the spectral acquisition equipment. With the development of deep learning, on-chip devices, and related technologies, fiber-optic F–P sensing demodulation schemes, when integrated with traditional demodulation methods, can accommodate increasingly complex application requirements, including integration, multi-parameterization (composite cavities), and in-situ compensation.

5 Applications in extreme scenarios

With the continued advancement of aerospace engineering, deep-earth energy exploration, the nuclear industry, and marine engineering, extreme environments, characterized by high temperature, high pressure, corrosive media, strong electromagnetic fields, and radiation, are imposing ever more stringent demands on sensor stability and reliability. Owing to their compact footprint, corrosion resistance, and inherent immunity to electromagnetic interference, fiber-optic sensors have emerged as a key technological solution for monitoring under such complex operating conditions.

The suitability of F–P sensors for various extreme environments depends not only on the target measurand, but also on the cavity architecture, fabrication method, and underlying sensing mechanism. Representative architectures include: (i) intrinsic in-fiber microcavities, typically fabricated by fusion splicing or femtosecond laser micromachining, which transduce external perturbations into changes in cavity length or refractive index; (ii) extrinsic cavities incorporating diaphragms or proof masses, where mechanical deformation under pressure or acceleration alters the air-gap length, offering high sensitivity, though thermal mismatch and fatigue remain concerns; (iii) MEMS-integrated cavities formed by silicon etching and wafer-level bonding, which benefit from batch fabrication and precise dimensional control, yet are constrained by the limited tolerance of silicon-based materials to ultra-high temperatures and aggressive corrosion; (iv) dual- or composite-cavity structures, which combine multiple cavities with distinct sensitivity and thermal coefficients, enabling simultaneous multiparameter measurement and effective cross-sensitivity decoupling; and (v) 3D-microprinted cavities fabricated via two-photon polymerization, which allow flexible three-dimensional geometries for customized sensing, but face challenges in material shrinkage, thermal stability, and reproducibility. These configurations operate through distinct transduction principles, whether changes in cavity length or refractive index, deformation of mechanical sensing elements, or differential responses among multiple cavities. Furthermore, the selection of materials, fabrication processes, and packaging schemes must be tailored to specific service conditions, including high temperature, high pressure, corrosion, strong electromagnetic fields, and radiation. The following sections therefore discuss the structural design and environmental adaptability of F–P sensors in representative extreme-environment applications.

5.1 Applications in high-temperature environments

Liu et al. (2026) proposed an all-silica fiber-optic F–P vibration sensor based on MEMS and laser welding. A beam-proof mass sensitive structure was fabricated using femtosecond laser. Combined with a gold-plated FBG and three-wavelength demodulation, temperature decoupling was achieved. The sensitivity at 800 °C was 0.0082 rad/g, and the accuracy after compensation was better than 2.31% F.S.¹⁰¹.

Li et al. (2026) developed a microsphere-lens-enhanced fiber-optic F–P high-temperature strain sensor using a cascaded structure of EFPI and GFBG. A microsphere lens was used to focus the optical field to maintain fringe contrast. The sensitivity at 800 °C was 22.87 nm/με, with an error < 2.17% F.S., and the sensor could withstand >4000 με (Fig. 10(a)-i)¹⁰².

Figure 10. High-temperature and high-pressure fiber-optic F–P sensing applications. (a) High-temperature environment application, (b) high-pressure environment application; (a) i Schematic diagram of an EFPI with microsphere -integrated optical fibers, (a) ii Schematic diagram of a sapphire-based fiber-optic EFPI sensor, (b) i All-silica high-temperature resistant fiber-optic F–P air pressure sensor, (b) ii hard-core diaphragm-enhanced all-fiber F–P high-pressure sensor. Figure reproduced from: (a) i ref¹⁰², Optica Publishing Group; (a)ii ref³⁹, Optica Publishing Group; (b) i ref¹⁰³, Wiley-VCH Verlag; (b) ii ref¹⁰⁵, IEEE.

Liao et al. (2025) developed a sapphire MEMS dual-cavity F-P pressure sensor. Three layers of sapphire were bonded to form a vacuum pressure cavity and a temperature measurement cavity. The dual cavities independently decoupled pressure and temperature. The resolution was 0.1% F.S. in the range of 0–1 MPa, and the pressure accuracy at 1500 °C was 0.86% F.S. (Fig. 10(a)-ii)³⁹.

5.2 Applications in high-pressure environments

Zhu et al. (2025) developed an all-silica high-temperature resistant fiber-optic F–P air pressure sensor. A quartz capillary and an optical fiber were CO₂ laser welded to form a sealed cavity. A cascaded femtosecond FBG was used to achieve in-situ temperature compensation. The sensitivity in the range of 0–14 MPa was 58 nm/MPa, with a room temperature error of 0.8% F.S. In the range of 25–655 °C, the sensor could measure 0–3.2 MPa, with a zero drift of 0.096% F.S. (Fig. 10(b)-i)¹⁰³.

He et al. (2023) proposed a quasi-distributed F–P high-pressure sensing scheme based on hollow-core photonic bandgap fibers. Combined with an optically carried microwave interferometry system, multi-node parallel demodulation was achieved. The measurement range was 0–100 MPa, with a pressure sensitivity of –25.6 pm/MPa, a temperature cross-sensitivity of only –0.02 MPa/°C, and a spatial resolution of centimeter level¹⁰⁴.

He et al. (2022) developed a hard-core diaphragm-enhanced all-fiber F–P high-pressure sensor. A high-germanium-doped multimode fiber and a single-mode fiber were spliced to form a cavity. The hard-core diaphragm multiplied the sensitivity, with a mechanical resonance frequency >100 kHz. The sensitivity in the range of 40–200 °C was 0.23 V/MPa, with hysteresis <1.5% and repeatability <0.8%. The results were consistent with those of a Kistler sensor (Fig. 10(b)-ii)¹⁰⁵.

Dai et al. (2022) proposed a suspended fiber-optic F–P high-pressure sensor. By adjusting the length ratio of the capillary to the air cavity (L₂/L₁), controllable sensitivity enhancement was achieved. When L₂/L₁ increased from 13.4 to 114.6, the sensitivity increased from –361.41 to –2975.21 pm/MPa. The measurement uncertainty in the range of 0–80 MPa was 0.03%, and the temperature crosstalk was 0.0011 MPa/°C¹⁰⁶.

5.3 Applications in liquid environments

Marine monitoring: Zhao et al. (2025) studied a silicon-based F–P interferometric temperature sensor. A single-crystal silicon chip was used as the sensitive chip, utilizing the high thermo-optic coefficient of silicon to achieve high-sensitivity temperature measurement without the need for coating. A current-modulated DFB laser combined with an HCN gas cell for wavelength calibration and a cross-correlation algorithm was used to improve stability¹⁰⁷. The TC4 titanium alloy pressure-resistant packaging withstands pressure up to 115 MPa, making it suitable for full-ocean-depth applications. Key specifications: sensitivity 75 pm/°C, resolution 6.2×10⁻⁵ °C, measurement range 0–35 °C. In the South China Sea trial, the standard deviation of the difference from the SBE37 CTD was <0.025 °C for the 0–3895 m profile, and the observation deviation at the 3900 m seabed over 87 days was only 0.0031 °C with no significant drift, demonstrating suitability for deep-sea profile observation, long-term seabed monitoring, and climate research (Fig. 11(a)-i).

Figure 11. Typical fiber-optic F–P sensing applications in extreme environments. (a) Liquid environment application, (b) strong electromagnetic interference environment application, (c) nuclear radiation environment application; (a)i ocean monitoring, (a)ii biomedical application, (a)iii humidity sensor, (b)i high-voltage power system application, (b)ii nuclear magnetic resonance imaging (MRI) application, (c)i nuclear power plant application, (c)ii nuclear reactor application, (c)iii nuclear waste monitoring application. Figure reproduced from: (a)i ref¹⁰⁷, IEEE; (a)ii ref¹⁰⁸, Elsevier; (a)iii ref¹⁰⁹, IEEE; (b)i ref¹¹⁰, IOP Science; (b)ii ref¹¹², MDPI; (c)i ref¹¹³, MDPI; (c)ii ref⁵⁴, MDPI; (c)iii ref¹¹⁴, Academic Press Inc.

Biomedicine: Wang et al. (2025) proposed a fiber-optic F–P cavity pH sensor based on a polyaniline deposited film. A polished fiber end face served as the cavity mirror, and a PANI sensitive film was deposited by in-situ oxidative polymerization. Reversible protonation/deprotonation changes the refractive index of the cavity to achieve pH response, with an in-fiber grating for temperature compensation¹⁰⁸. The sensor dimensions are 40 mm × 12 mm, with a measurement range of pH 2–12. The linearity at 0 °C is R² = 0.98838, the equivalent cavity length sensitivity is −1.228 μm/pH, and the wavelength sensitivity is −0.28682 nm/pH. It exhibits good linear response at low temperature and is suitable for medical fluid detection, biological fermentation, and in-situ monitoring of chemical processes (Fig. 11(a)-ii).

Humidity detection: Yang et al. (2026) developed an ultra-sensitive fiber-optic F-P humidity sensor based on a PVA thin film and a note-shaped open cavity structure. The interference cavity directly interacts with the external environment through a cold splice¹⁰⁹. In the range of 6.12%–16.69% RH, the wavelength sensitivity reaches 60.25 nm/%RH, effectively suppressing temperature interference. This provides a high-performance, electromagnetic interference immune sensing solution for online monitoring of trace moisture in high-voltage power equipment such as GIS (Fig. 11(a)-iii).

5.4 Applications in strong electromagnetic interference environments

High-voltage power systems: Zhang et al.(2022) studied a fiber-optic sensing system based on F–P interference. A charge induction probe and a piezoelectric thin film were used to drive the F–P cavity length change, achieving nanosecond-level partial discharge signal detection¹¹⁰. The signal-to-noise ratio is 18.73–19.29 dB, and the average sensitivity is 17.2% higher than that of the UHF antenna method. The system is suitable for online partial discharge monitoring and early fault warning of transformers, GIS, switchgears, and other equipment (Fig. 11(b)-i).

Radar stations: Hossain et al. (2023) used an FPI to observe thermospheric OI 630.0 nm nightglow emission and retrieved neutral temperature and wind fields at altitudes of 220–280 km¹¹¹. For the first time, seasonal variations of thermospheric temperature and wind fields during solar minimum in the Indian equatorial region were obtained. The temperature ranged from 601 to 849 K. The meridional wind agreed with the HWM14 model, and the zonal wind revealed the PRE effect and regional differences, providing ground-based observational support for ionospheric coupling studies and atmospheric model validation.

Nuclear magnetic resonance equipment: Hwang et al. (2020) proposed a dual-layer metallic grating polarizing unit based on an F-P resonance. A dielectric layer between two layers of gold nano gratings forms an F–P cavity, enhancing TM polarization transmission and suppressing TE polarization¹¹². At a wavelength of 4.5 μm, the transmittance is 0.49, and the polarization extinction ratio is 132. The extinction ratio variation is less than 6% over an incident angle range of 0°–25°. The device is suitable for DOFP polarization imaging systems (Fig. 11(b)-ii).

5.5 Applications in nuclear radiation environments

Nuclear power plants: Pang et al. (2026) proposed a sapphire MEMS fiber-optic F–P vibration sensor with all-ceramic packaging using alumina ceramic. A sapphire composite F–P cavity was fabricated by UV laser micromachining¹¹³. The sensor operates stably from −55 to 1200 °C, with a room temperature sensitivity of 20.12 nm/g and a sensitivity of 24.89 nm/g at 1200 °C. The resonant frequency is ~3300 Hz, and the bandwidth is 100–1000 Hz. This breaks through the bottleneck of long-term operation of silicon/quartz-based sensors at 800 °C, making it suitable for high-temperature vibration monitoring of the primary circuit and nuclear main pump in nuclear power plants (Fig. 11(c)-i).

Nuclear reactors: We proposed a fiber-optic F-P strain sensing system based on non-scanning correlation demodulation. A double-spring-leaf ring structure transfers the strain of the fuel plate to the F–P cavity⁵⁴. The sensitivity at 300 °C is 12.6 nm/με, with a dynamic response of 10–500 Hz. The consistency deviation of thermal-hydraulic experimental data is less than 1.5%. Temperature cross-interference is eliminated by cavity length calibration. The system is used for structural health monitoring of pressurized water reactor fuel assemblies (Fig. 11(c)-ii).

Nuclear waste treatment: Gao et al. (2025) proposed a dual-FPI Vernier-effect MOF-functionalized fiber-optic sensor using a UiO-66-PYDC metal-organic framework for specific adsorption of iodine vapor¹¹⁴. The Vernier cascade sensitivity is 77.76 pm/ppb, and the detection limit is 10 ppb. Complementary cancellation of the polymer TOC and TEC suppresses the temperature cross-sensitivity to −3.35 pm/°C. The sensor enables trace iodine monitoring from 0 to 80 °C and is suitable for safety monitoring of radioactive iodine off-gas in the nuclear industry (Fig. 11(c)-iii).

5.6 Applications in other extreme environments

Fiber-optic F–P sensors exhibit broad application potential across interdisciplinary fields. Zhang et al.(2022) fabricated an all-fiber diaphragm-type EFPI using femtosecond laser, achieving linear measurement of 0–7 MPa at an ultra-low temperature of −196 °C, with pressure sensitivities of 111.17 nm/MPa (increasing pressure) and 111.22 nm/MPa (decreasing pressure)¹¹⁵. Li et al. (2024) proposed a dynamic spectral demodulation method based on EFPI, detecting vibration signals as low as 6 Hz by Fourier sideband characteristics, with a sideband sensitivity of approximately 17 mW/g, suitable for earthquake and building health monitoring¹¹⁶. Amoudry et al. (2020) employed a movable D-shaped dielectric mirror in a high-power optical resonant cavity to suppress thermoelastic higher-order transverse mode degeneracy, achieving stable operation at an average intracavity power of 200 kW with a fundamental mode loss of only about 8 ppm¹¹⁷. Xie et al. (2022) designed a VO₂/BaF₂ stacked F–P multilayer film, utilizing multiple resonances to achieve a mid-infrared normal emissivity close to 0.95 for spacecraft thermal control coatings, with a tunability of 0.78¹¹⁸. Allam et al. (2025) combined a porous silicon F-P interferometer with a molecularly imprinted polymer to achieve selective detection of ticagrelor, with a linear range of 1.0×10⁻⁷–5.0×10⁻⁵ M and a detection limit of 79 nM, which has been applied to pharmaceutical formulations and human plasma analysis¹¹⁹.

6 Future development

Fiber-optic F–P sensing systems are transitioning from a laboratory paradigm of "discrete devices + benchtop equipment + manual algorithms" toward an engineering paradigm characterized by integration, intelligence, and networking. At the physical sensing layer, optical fibers and probes are evolving toward micro/nano fabrication, new material integration, and multi-parameter integration. Spectral readout technology is developing toward chip-scale, high-speed swept-frequency, and spectrometer-free demodulation. At the demodulation level, intelligent algorithms such as deep learning and compressed sensing enhance noise immunity and speed. At the system level, hybrid multiplexing, weak-reflector grating arrays, and on-chip optical switching support large-capacity, low-crosstalk quasi-distributed networking. Ultimately, these technologies rely on low-cost packaging, standardized interfaces, and optoelectronic integrated chips to achieve large-scale deployment in scenarios such as power, medical, and structural monitoring, promoting fiber-optic F–P sensing from the laboratory to industrial sites and daily life. These directions are mutually reinforcing and collectively constitute the frontier research landscape of this field.

(1) Specialty optical fibers and integrated probes

Despite their superior high-temperature and corrosion resistance, sapphire fibers still face practical limitations in terms of multimode transmission, high-temperature cladding, fiber coupling and packaging, and fabrication consistency. The relatively large core diameter and weak optical confinement of conventional sapphire fibers tend to excite higher-order modes, which cause spectral distortion and degrade demodulation stability. Furthermore, the thermomechanical mismatch between sapphire and silica fibers poses additional challenges for high-temperature coupling and packaging. To advance the engineering application of sapphire-fiber sensors, it is therefore critical to improve mode control, thermal compatibility, fabrication consistency, and long-term reliability.

Using single-crystal sapphire fiber as the core matrix, three major trends are emerging: First, evolving from unclad sapphire to high-temperature cladding systems that are stable above 1200 °C, such as spinel, nanoporous alumina, and ion-implanted modified layers, to solve multimode interference and interface failure. An example is single-crystal sapphire fiber fabricated by the LHPG method (Fig. 12(a)-i). Second, developing new fiber materials such as YAG, zirconia-based, and high-melting-point ceramics to extend applicability to ultra-high temperatures of 1400–1800 °C and strong acid/alkali environments, e.g., YAG-clad single-crystal fiber prepared by sol-gel coating. Third, balancing high-temperature resistance, low loss, and single-mode transmission through microstructured design, such as windmill-shaped microstructured sapphire fiber and negative-curvature hollow-core anti-resonant fiber. The latter uses an air core to greatly reduce light-material contact, offering strong corrosion resistance and high-temperature tolerance (Fig. 12(a)-ii)^120,121.

Figure 12. Future development of fiber-optic F–P sensing measurement systems. (a) Development of physical component structures, (b) miniaturization development of spectral reconstruction technology, (c) development of intelligent algorithms; (a)i EFG high-temperature resistant optical fiber, (a)ii special hollow-core fiber structure, (a)iii SMF-28 fiber end-face F-P structure, (a)iv laser-processed fiber end-face, (b)i MM-WBG on-chip spectrometer, (b)ii MRR on-chip spectrometer, (c)i ADPNet decoupling algorithm, (c)ii CNN+LSTM demodulation algorithm. Figure reproduced from: (a)i ref¹²⁰, AIP Publishing; (a)ii ref¹²¹, IOP Science; (a)iii ref¹²³, MDPI; (a)iv ref¹²², Elsevier; (b)i ref¹²⁵, Optica Publishing Group; (b)ii ref¹²⁶, Wiley Online Library; (c)i ref¹²⁷, IEEE Xplore; (c)ii ref¹³², IEEE;

Fiber-optic F–P sensing probes are moving from discrete assembly to integrated fabrication on the fiber end face. Sensitive microcavities, elastic diaphragms, and supporting structures are formed in situ at the fiber tip, achieving adhesive-free, packaging-free, solder-joint-free all-optical passive integration, eliminating interfacial thermal stress and high-temperature drift. Probes achieve homogeneous integration using high-temperature resistant materials such as sapphire, MgO, SiC, and high-purity silica, with operating temperatures exceeding 1200 °C (Fig. 12(a)-iii). In fabrication, end-face 3D printing by two-photon polymerization femtosecond laser direct writing can directly shape complex structures such as F–P microcavities and cantilevers, followed by high-temperature degreasing and sintering to achieve ceramization, replacing traditional multi-step MEMS processes and improving consistency and batch production capability (Fig. 12(a)-iv). However, the scalability and reproducibility of F-P sensors remain sensitive to fabrication tolerances. In MEMS fabrication, variations in diaphragm thickness, etching depth, and bonding gap can alter the initial cavity length and mechanical sensitivity. Laser micromachining may introduce cavity-dimensional variations due to fluctuations in laser energy and positioning accuracy, while its point-by-point processing nature limits fabrication throughput. For 3D-microprinted cavities, material shrinkage can further compromise geometric reproducibility. Therefore, stable process control and standardized batch fabrication remain critical for large-scale manufacturing. Multi-FP composite cavity probes can achieve simultaneous measurement of temperature-pressure-strain. For example, an all-ceramic MgO dual-FP cavity probe combined with frequency-domain separation and cross-correlation demodulation achieves crosstalk-free decoupling, meeting the requirements of multi-physics coupling monitoring^122–124.

(2) Miniaturization development of spectral reconstruction

On-chip spectrometers, offering core advantages such as miniaturization, low power consumption, CMOS compatibility, and ease of networking and integration, are key components of miniature passive optical sensing systems. They support applications including environmental monitoring, biosensing, and spaceborne remote sensing. The mainstream approaches are divided into three categories: multimode grating filter type, micro-ring resonant computational type, and meta surface speckle-based type, with ongoing breakthroughs in resolution, bandwidth, and integration density.

In 2025, we proposed a multimode waveguide Bragg grating spectrometer based on an SOI platform. Four thermo-optically tunable MM-WBG filter arrays achieve a bandwidth of 35 nm and a resolution of <150 pm, with a single-channel linewidth as low as 40 pm, and a spiral structure improves space utilization (Fig. 12(b)-i). In 2023, we proposed a micro-ring resonator array computational spectrometer based on a silicon nitride platform. Three micro-rings combined with a convex optimization algorithm solve underdetermined equations, requiring no external tuning, achieving a bandwidth of >12 nm and a resolution of <0.17 nm, suitable for passive, low-power scenarios (Fig. 12(b)-ii). Zhang et al. (2025) proposed a three-layer cascaded disordered meta surface speckle spectrometer. Wavelength-dependent speckles are generated by in-plane diffraction and read out by two-dimensional imaging. On a 150 μm × 950 μm chip, a bandwidth of 100 nm and a resolution of 70 pm are achieved, with a spectral channel density of 10021 ch/mm², breaking through the channel density bottleneck of traditional structures^125–127.

(3) Complex physical quantity decoupling and intelligent demodulation algorithms

Traditional decoupling of multi-field coupling in optical fiber sensing relies on sensitivity matrices and polynomial fitting, suffering from large crosstalk and poor adaptability. Artificial intelligence, with its end-to-end learning and nonlinear mapping capabilities, provides a new path for adaptive decoupling of multi-field interference such as temperature-pressure and temperature-strain. A DenseNet dual-branch network achieves EFPI-FBG temperature pressure decoupling with a temperature MAE of 0.907 °C and a latency of only 0.38 ms¹⁰⁰. A categorical boosting tree algorithm completes three-parameter decoupling of FBG-FPI with a temperature RMSE as low as 0.026 °C¹²⁸. The ADPNet network achieves strain decoupling with an RMSE of 141.755 με, better than traditional compensation methods¹²⁹. The deep integration of artificial intelligence and optical fiber sensing has become a core path to solve multi-field coupling interference and improve demodulation accuracy and robustness¹³⁰. AI achieves performance upgrades in all scenarios from point to distributed sensing through signal denoising, cross-sensitivity compensation, and accelerated computation. For F–P sensing, federated learning demodulation achieves an accuracy of ±0.005 nm while protecting data privacy¹³¹. CNN-LSTM achieves direct F–P strain demodulation with an RMSE as low as 0.524 με¹³². Emerging trends include: (a) adaptive dynamic decoupling incorporating online incremental learning and weakly supervised training; (b) high-dimensional, multi-field unified decoupling that enables collaborative demodulation of multiple parameters (e.g., strain, bending, vibration); and (c) mechanism-data fusion decoupling aimed at improving model interpretability and generalization. Consequently, artificial intelligence is emerging as a core facilitator of adaptive decoupling in multi-field coupling scenarios, steering the development of optical fiber sensing toward high-precision and intelligent systems.

(4) Miniaturized passive networking technologies

In miniaturized passive optical networking sensing terminals, FMCW interferometry combined with the virtual Vernier effect, while intensity-modulated fiber optic sensing (IM-FOS), leverages the advantages of low cost, high integration, and ease of multiplexing to achieve complementary breakthroughs in high-precision distributed temperature measurement and industrial multi-parameter sensing, respectively.

Mo et al. (2025) proposed a quasi-distributed temperature measurement system based on FMCW interferometry, utilizing software-generated artificial reference spectra to achieve the virtual Vernier effect for sensitivity magnification. With a sensing distance of 82.3 m and a spatial resolution of 23.9 μm, the system achieved a maximum sensitivity of −280.67 pm/℃ in the range of 30–70 °C, a magnification factor of 6.13 times, and linearity >99.8% (Fig. 13(a)-i). Malinka et al. (2025) reviewed industrial-grade IM-FOS technologies, covering six categories including macrobending, microbending, and evanescent field, capable of multi-parameter detection such as displacement, pressure, and temperature. The bending resolution reaches 0.01°, and the evanescent field sensitivity reaches 154 dB/RIU. The technology supports various multiplexing schemes for high-density networking, with outstanding potential for chip-scale integration and IoT integration (Fig. 13(a)-iii)^133,134.

Figure 13. Engineering-scale development of fiber-optic F–P sensing technology. (a) Miniaturized networking technology, (b) current products and engineering applications; (a)i multi-core fiber vector angular displacement monitoring, (a)ii sensor applied to metal corrosion, (a)iii multi-point temperature detection based on FMCW interferometry; (b)i fiber-optic sensing demodulation products, (b)ii F–P sensor applications in bridges and thermal power plants. Figure reproduced from: (a)i ref¹³⁴, Nanomaterials; (a)ii ref¹³⁴, Nanomaterials; (a)iii ref¹³³, iScience.

(5) Engineering and large-scale applications

The key technologies for transitioning fiber-optic F-P sensing from the laboratory to engineering field applications are still in a critical development period. Mature equipment systems represented by the HYPERION si255 and sm125 interrogators and the ENLIGHT software have formed a well-established ecosystem: the si255 supports 16 channels, a scan rate of 5000 Hz, and a bandwidth of 160 nm, capable of simultaneously measuring nearly a thousand sensors, meeting high-dynamic requirements such as aerospace wind tunnel testing; the sm125 is based on a tunable F–P filter and serves long-term static monitoring of bridges, dams, oil wells, etc., with a compact modular design—over half of the fiber optic sensors worldwide adopt this series; ENLIGHT provides an integrated platform from configuration and data acquisition to analysis and alarm archiving, compatible with multi-language APIs such as Python and MATLAB. These equipment systems have been applied to major projects such as rock mass monitoring at Mount Rushmore in the United States, oil and gas pipelines, offshore platforms, aircraft composite material testing, and temperature measurement of power switchgear, strongly supporting the development of industry standards from the product level to the system level^135,136.

In industrial applications, fiber-optic F–P sensors have been successfully deployed for continuous pressure monitoring in oil and gas wells. Their high-temperature, high-pressure reliability is ensured by a metal–quartz tube mechanical pressure seal combined with a laser-based, adhesive-free hot-melt sealing technique¹³⁷. In aero-engine dynamic strain measurement, the F–P system based on phase-tracking demodulation achieves a sensitivity of 100.955 pm/με (approximately 100 times that of an FBG), and a dynamic response frequency of 10 kHz¹³⁸.

Silicon photonic integrated demodulation schemes, such as AWG and MZI, have demonstrated picometer-level accuracy in FBG interrogation, offering a promising reference for chip-scale demodulation of F-P systems^139,140. Equipped with an embedded AI/ML engine, the VIAVI FTH-DAS interrogator enables real-time event classification and true physical quantity measurement, effectively transforming the device from a mere data acquisition unit into an intelligent terminal¹⁴¹.

Although miniaturization, integration, and intelligent sensing represent important development directions for F–P sensing systems, their large-scale industrial deployment continues to face challenges in long-term reliability, manufacturing consistency, and standardized evaluation. In extreme environments, long-term drift and repeated temperature-pressure cycling may compromise measurement stability, while fabrication tolerances in miniature F–P cavities and packaging structures can lead to significant device-to-device performance variations. Moreover, on-chip and intelligent demodulation schemes still require careful trade-offs among resolution, bandwidth, computational resources, model generalization, and cost. Therefore, long-term reliability, batch-to-batch consistency, and system-level engineering capability remain the key issues to be resolved for the industrialization of F–P sensing technologies.

7 Conclusion

Fiber-optic F–P sensors, by virtue of their high sensitivity, immunity to electromagnetic interference, resistance to high temperature and corrosion, ease of miniaturization, and long-distance transmission capability, have become an important technological approach for intelligent sensing in extreme environments. They are evolving from single-parameter measurement toward multi-parameter collaborative sensing systems. Structurally, F–P sensors are transitioning from simple interference cavities toward miniaturized, chip-based, and integrated designs. Technologies such as femtosecond laser processing, MEMS micromachining, and 3D micro-printing enable the direct fabrication of complex micro-structured sensitive units on fiber end faces, achieving highly sensitive detection of micro-force, displacement, vibration, and acoustics. New materials such as sapphire and hollow-core fibers further expand the measurement range for high temperature, high pressure, gas, humidity, and magnetic fields. Moreover, composite cavities and cascaded F–P/FBG structures endow sensors with multi-parameter synchronous measurement capability and in-situ temperature compensation, thereby promoting their engineering applications.

In terms of demodulation, three major technical routes have been established: intensity-based, phase-based, and spectral-based methods. For intensity demodulation, harmonic amplitude ratio, dual-wavelength ratio, and active feedback locking have improved demodulation stability. Phase demodulation techniques, including phase-generated carrier (PGC), multi-wavelength quadrature phase shift, and non-scanning correlation demodulation have achieved high-resolution, wide-dynamic-range detection. Broadband spectral demodulation methods, such as FFT, cross-correlation, and spectral comparison have improved absolute cavity length accuracy and composite cavity decoupling capability. Furthermore, artificial intelligence and deep learning, by leveraging neural networks to establish nonlinear mapping from spectra to physical quantities, provide a new approach for adaptive demodulation.

In terms of performance, sensors have achieved nm-level displacement detection and mPa-level acoustic pressure detection. Some sapphire MEMS structures can achieve ultra-high-temperature pressure measurement up to 1500 °C and high-pressure monitoring up to 100 MPa, with dynamic response capability reaching the kHz to MHz level. However, challenges remain regarding long-term stability, packaging thermal mismatch, material reliability, and multi-physics cross-coupling.

In practical applications, F–P sensors have been deployed in engineering scenarios such as aero-engines, gas turbines, nuclear power, deep sea, oil and gas, liquid metals, and intense radiation fields. By virtue of all-metal packaging, sapphire structures, and non-electrical measurement characteristics, they demonstrate advantages that conventional electrical sensors cannot easily replace. They also show promising potential in emerging directions including ultrasonic detection, photoacoustic imaging, gas photothermal detection, and miniaturized biomedical sensing.

Future development focuses on the following aspects: (1) Improving reliability and long-term stability in extreme environments through sapphire fibers, hollow-core fibers, ceramics, and all-metal packaging to enhance environmental adaptability. (2) Developing multi-cavity, multi-modal sensitive units and AI-assisted decoupling algorithms, combined with digital twins, to achieve real-time decoupling and intelligent compensation of multi-physical quantities. (3) Promoting the integration of MEMS, on-chip spectrometers, and integrated photonic chips to construct miniaturized intelligent sensing nodes integrating light source, interference cavity, demodulation, and wireless communication. (4) Leveraging FPGA, high-speed DAQ, and edge AI to achieve high-speed real-time demodulation and online anomaly diagnosis. (5) Establishing standardized testing, modular packaging, and reliability verification systems to promote large-scale applications in nuclear industry, aerospace, energy equipment, intelligent manufacturing, and other fields.

Overall, fiber-optic F–P sensing technology is transforming from high-sensitivity laboratory devices toward intelligent sensing systems for complex environments. It is expected to play a key role in extreme environment monitoring, major equipment health management, and intelligent perception networks, thereby becoming an important development direction for a new generation of high-end sensing technologies.

Competing interests

The authors declare no competing interests.

Article History

Received Date: 2026-05-28

Accepted Date: 2026-09-04

Available Online: 2026-09-28

Corresponding Author

• Jie Zhang, zhangjie@cqu.edu.cn

• Issue Info: Vol. 1, No. 1, 260004-1–260004-31

• DOI: 10.67704/technol.2026.260004

Citation

Wang N, Kang W, Tan J et al. Fiber-optic Fabry–Perot sensors and their demodulation techniques for extreme environment monitoring. Technology 1, 260004 (2026)

Author Information

Ning Wang

The Key Laboratory of Optoelectronic Technology & System, Education Ministry of China, Chongqing University, Chongqing 400044, China

Weijian Kang

The Key Laboratory of Optoelectronic Technology & System, Education Ministry of China, Chongqing University, Chongqing 400044, China

Jiahang Tan

The Key Laboratory of Optoelectronic Technology & System, Education Ministry of China, Chongqing University, Chongqing 400044, China

Feng Qin

The Key Laboratory of Optoelectronic Technology & System, Education Ministry of China, Chongqing University, Chongqing 400044, China

Yong Zhu

The Key Laboratory of Optoelectronic Technology & System, Education Ministry of China, Chongqing University, Chongqing 400044, China

Corresponding author Jie Zhang, E-mail: zhangjie@cqu.edu.cn

The Key Laboratory of Optoelectronic Technology & System, Education Ministry of China, Chongqing University, Chongqing 400044, China

Copyright & License

© The Author(s) 2026. Published by Technology Publishing.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

References

1. 1. Liu Q, Wang Y, Zhao FJ et al. A review of the research progress of sensor monitoring technology in harsh engineering environments. Sensors 25, 6308 (2025). DOI: 10.3390/s25206308

2. 2. Meena JS, Xiang LLY, Kannan DD et al. Extreme-temperature resistant, flexible, and sensitive strain sensor for aerospace parachute deployment. Adv Mater Technol 11, e02467 (2026). DOI: 10.1002/admt.202502467

3. 3. Jahanshahi H, Zhu ZH. A comprehensive review of tactile sensing technologies in space robotics. Chin J Aeronaut 38, 103423 (2025). DOI: 10.1016/j.cja.2025.103423

4. 4. Liu YH, Wang JL, Liu T et al. Triboelectric tactile sensor for pressure and temperature sensing in high-temperature applications. Nat Commun 16, 383 (2025). DOI: 10.1038/s41467-024-55771-0

5. 5. Chen HC, Kuo CW, Chang TC et al. Investigation of the capacitance-voltage electrical characteristics of thin-film transistors caused by hydrogen diffusion under negative bias stress in a moist environment. ACS Appl Mater Interfaces 11, 40196–40203 (2019). DOI: 10.1021/acsami.9b11637

6. 6. Taniker S, Costanza V, Celli P et al. Capacitive temperature sensing via displacement amplification. IEEE Sens J 22, 10388–10395 (2022). DOI: 10.1109/JSEN.2022.3169585

7. 7. Hu YJ, Huang YP, Yang ZQ et al. Optimization of optical fiber monitoring technology and its rock pressure early warning mechanism in deep mining development environments. Eng Res Express 8, 035115 (2026). DOI: 10.1088/2631-8695/ae429c

8. 8. Wang PJ, Pan YF, Zhang JS et al. Miniaturized and highly sensitive fiber-optic Fabry-Perot sensor for mHz infrasound detection. Photonics Res 12, 969–978 (2024). DOI: 10.1364/PRJ.519946

9. 9. De Pauw B, Lamberti A, Ertveldt J et al. Vibration monitoring using fiber optic sensors in a lead-bismuth eutectic cooled nuclear fuel assembly. Sensors 16, 571 (2016). DOI: 10.3390/s16040571

10. 10. Wang YJ, Liu J, Yang N et al. Fabrication of waveguide Bragg gratings on sapphire chip and their temperature sensing characteristics. Chin J Lasers 52, 1202403 (2025). DOI: 10.3788/CJL250466

11. 11. Zhang XP, Zhang YX, Wang LA et al. Current status and future of research and applications for distributed fiber optic sensing technology. Acta Opt Sin 44, 0106001 (2024). DOI: 10.3788/AOS231473

12. 12. Wang N, Zhu Y, Zhang J. Research status of fiber Fabry-Perot sensing technology under high temperature and high pressure environment. Laser Optoelectron Prog 60, 1106005 (2023). DOI: 10.3788/LOP230722

13. 13. Li TL, Wu DJ, Khyam MO et al. Recent advances and tendencies regarding fiber optic sensors for deformation measurement: a review. IEEE Sens J 22, 2962–2973 (2022). DOI: 10.1109/JSEN.2021.3138091

14. 14. Kumar V, Raghuwanshi SK, Kumar S. Advances in SPR-based fiber optic sensors for voltage/electric field measurement - A review. IEEE Sens J 24, 11986–11997 (2024). DOI: 10.1109/JSEN.2024.3373389

15. 15. Zhou CX, Jia ZK, Song SB et al. Application of FBG sensor in health monitoring of engineering building structure: a review. Sens Rev 45, 129–145 (2025). DOI: 10.1108/SR-06-2024-0572

16. 16. Wang M, Dai SX, Zhang PQ et al. Research progress on evanescent wave sensor based on mid-infrared optical fiber. J Chin Ceram Soc 50, 1117–1131 (2022). DOI: 10.14062/j.issn.0454-5648.20211175

17. 17. Yu SJ, Cao K, Cao JJ et al. Portable SPR sensor with laser interference lithography – Fabricated Ag nanogratings for high-sensitivity biosensing. Sens Actuators A Phys 399, 117399 (2026). DOI: 10.1016/j.sna.2025.117399

18. 18. Tong RJ, Zhao KJ, Xing B et al. An optical fiber sensor for salinity and temperature simultaneous detection based on dual SPR effect. Opt Laser Technol 175, 110760 (2024). DOI: 10.1016/j.optlastec.2024.110760

19. 19. Zhang WJ, Su D, Qiao XG. High-temperature and high-pressure sensor based on sensitivity-enhanced suspended all-optical fiber Fabry-Pérot interferometer for downhole oil applications. IEEE Trans Instrum Meas 74, 9518709 (2025). DOI: 10.1109/TIM.2025.3562996

20. 20. Hu Y, Wei HM, Ma ZW et al. Microbubble-based optical fiber Fabry-Perot sensor for simultaneous high-pressure and high-temperature sensing. Opt Express 30, 33639–33651 (2022). DOI: 10.1364/OE.465315

21. 21. Islam MR, Ali MM, Lai MH et al. Chronology of Fabry-Perot interferometer fiber-optic sensors and their applications: a review. Sensors 14, 7451–7488 (2014). DOI: 10.3390/s140407451

22. 22. Huang YW, Tao J, Huang XG. Research progress on F-P interference-based fiber-optic sensors. Sensors 16, 1424 (2016). DOI: 10.3390/s16091424

23. 23. Cao XN, Yang H, Wu ZL et al. Ultrasound sensing with optical microcavities. Light Sci Appl 13, 159 (2024). DOI: 10.1038/s41377-024-01480-8

24. 24. Huynh NT, Zhang E, Francies O et al. A fast all-optical 3D photoacoustic scanner for clinical vascular imaging. Nat Biomed Eng 9, 638–655 (2025). DOI: 10.1038/s41551-024-01247-x

25. 25. Zhang L, Shang XG, Cao SM et al. Optical steelyard: high-resolution and wide-range refractive index sensing by synergizing Fabry–Perot interferometer with metafibers. PhotoniX 5, 24 (2024). DOI: 10.1186/s43074-024-00138-3

26. 26. Xu YF, Teng HB, Zhao XY et al. Fiber-tip Fabry-Perot interferometric photothermal flexible gas sensor with capillary microcavity. ACS Sens 11, 3986–3992 (2026). DOI: 10.1021/acssensors.6c00294

27. 27. Li CX, Liu YQ, Han DD et al. A tactile gripper on an optical fiber for perception and actuation. Nat Commun 16, 11513 (2025). DOI: 10.1038/s41467-025-66575-1

28. 28. Fan XH, Li LY, Li ZJ et al. Monolithic 3D micro-encapsulation on optical fiber tips via femtosecond laser direct writing. Int J Extrem Manuf 8, 045004 (2026). DOI: 10.1088/2631-7990/ae4e6f

29. 29. Lu QB, Wang YA, Wang XX et al. Review of micromachined optical accelerometers: from mg to sub-μg. Opto-Electron Adv 4, 200045 (2021). DOI: 10.29026/oea.2021.200045

30. 30. Zhao MH, Qi YH, Wang HL et al. Optical interferometric MEMS accelerometers. Laser Photonics Rev 18, 2300713 (2024). DOI: 10.1002/lpor.202300713

31. 31. Wang YH, Chen BT, Wu GR et al. Fiber-optic pressure sensors: recent advances in sensing mechanisms, fabrication technologies, and multidisciplinary applications. Sensors 25, 6336 (2025). DOI: 10.3390/s25206336

32. 32. Chen Y, Lu DQ, Xing H et al. Recent progress in MEMS fiber-optic Fabry-Perot pressure sensors. Sensors 24, 1079 (2024). DOI: 10.3390/s24041079

33. 33. Li TL, Zhang J, Al-Hadad M et al. Recent progress in fiber-optic acoustic sensor and its applications: a review. IEEE Sens J 24, 25249–25260 (2024). DOI: 10.1109/JSEN.2024.3422186

34. 34. Shang XG, Wang N, Cao SM et al. Fiber-integrated force sensor using 3D printed spring-composed Fabry-Perot cavities with a high precision down to tens of piconewton. Adv Mater 36, 2305121 (2024). DOI: 10.1002/adma.202305121

35. 35. Li TY, Liu Y, Xu WJ et al. Light control of a movable microbubble in an ethanol-filled fiber microcavity for displacement measurement. Opt Lett 49, 790–793 (2024). DOI: 10.1364/OL.510306

36. 36. Cao SJ, Bai JD, Zheng YQ et al. High-temperature fiber-optic vibration sensor based on an atomic frequency standard. ACS Photonics 11, 3713–3721 (2024). DOI: 10.1021/acsphotonics.4c00885

37. 37. Wang P, Li TG, Lin H et al. Miniature optical fiber accelerometer based on an in-situ 3D microprinted ferrule-top Fabry-Pérot microinterferometer. Light Adv Manuf 6, 152–160 (2025). DOI: 10.37188/lam.2025.018

38. 38. Zhen YQ, Tu XT, Zhu JJ et al. Atomically smooth gold microflake-enabled fiber-tip Fabry–Perot interferometer for temperature and pressure sensing. ACS Appl Mater Interfaces 15, 35161–35169 (2023). DOI: 10.1021/acsami.3c04809

39. 39. Liao YK, Liu J, Dai YT et al. Temperature-compensated fiber-optic Fabry-Perot pressure sensor based on sapphire MEMS technology for high temperature environment up to 1500°C. Opt Express 33, 19453–19463 (2025). DOI: 10.1364/OE.560263

40. 40. Wei T, Zhang YQ, Zhang LQ et al. Research on ultra-high sensitivity fiber-optic cascaded Fabry-Perot resonator optical sensor based on the comb-spectrum vernier effect. J Lightwave Technol 42, 6182–6187 (2024). DOI: 10.1109/JLT.2024.3404783

41. 41. Shen YF, Jiang Y, Xie SR et al. Cascaded Fabry-Perot cavity and fiber Bragg grating on sapphire fibers for high-temperature strain sensing. Opt Lasers Eng 184, 108674 (2025). DOI: 10.1016/j.optlaseng.2024.108674

42. 42. Wang AQ, Fan XH, Chen G et al. Ultrahigh-sensitivity 3D-printed hollow Fabry-Pérot fiber ultrasound sensor for photoacoustic imaging. ACS Photonics 12, 4553–4562 (2025). DOI: 10.1021/acsphotonics.5c00944

43. 43. Zhao PC, Krishnaiah KV, Guo LH et al. Ultraminiature optical fiber-tip 3D-microprinted photothermal interferometric gas sensors. Laser Photonics Rev 18, 2301285 (2024). DOI: 10.1002/lpor.202301285

44. 44. Zhao ZK, Qian CW, Ni WJ et al. Microscale intrinsic fiber mode vernier photothermal spectroscopy gas sensing. Laser Photonics Rev 20, e02114 (2026). DOI: 10.1002/lpor.202502114

45. 45. Xiong C, Wang CY, Yu RW et al. 3D printed multicore fiber-tip discriminative sensor for magnetic field and temperature measurements. Light Adv Manuf 5, 84–94 (2024). DOI: 10.37188/lam.2024.018

46. 46. Chen MQ, Liu SY, Zhang C et al. Two-photon 3D printing FP microcavity sensor for simultaneous measurement of temperature and non-contact pressure. Opt Laser Technol 179, 111406 (2024). DOI: 10.1016/j.optlastec.2024.111406

47. 47. Li WT, Li C, Wan Z et al. Photothermally actuated Ti₃C₂T_x MXene-GO resonator for highly sensitive humidity sensing. Sens Actuators B Chem 461, 139983 (2026). DOI: 10.1016/j.snb.2026.139983

48. 48. Qin F, Zheng TT, Tan JH et al. Design and experiment of high temperature Fabry-Perot acceleration sensor with composite cavity. Opt Precis Eng 32, 3174–3183 (2024). DOI: 10.37188/OPE.20243221.3174

49. 49. Zhu C, Gerald RE, Huang J. Progress toward sapphire optical fiber sensors for high-temperature applications. IEEE Trans Instrum Meas 69, 8639–8655 (2020). DOI: 10.1109/TIM.2020.3024462

50. 50. Deng YQ, Jiang J. Optical fiber sensors in extreme temperature and radiation environments: a review. IEEE Sens J 22, 13811–13834 (2022). DOI: 10.1109/JSEN.2022.3181949

51. 51. Ma SN, Xu YP, Pang YX et al. Optical fiber sensors for high-temperature monitoring: a review. Sensors 22, 5722 (2022). DOI: 10.3390/s22155722

52. 52. Huang PJ, Wang N, Li JY et al. Fiber Fabry-Perot force sensor with small volume and high performance for assessing fretting damage of steam generator tubes. Sensors 17, 2899 (2017). DOI: 10.3390/s17122899

53. 53. Lin C, Wang N, Chen JJ et al. Study on fiber Fabry–Perot displacement sensor for vibration monitoring of fuel pin. Microw Opt Technol Lett 66, e34007 (2024). DOI: 10.1002/mop.34007

54. 54. Jiao JA, Chen JJ, Wang N et al. Study of a fiber optic Fabry-Perot strain sensor for fuel assembly strain detection. Sensors 22, 9097 (2022). DOI: 10.3390/s22239097

55. 55. Liu LL, Qin F, Xu YG et al. All-metal packaged temperature compensation fiber optic Fabry-Pérot strain sensor for high-temperature liquid metal environments. Measurement 256, 118501 (2025). DOI: 10.1016/j.measurement.2025.118501

56. 56. Tan JH, Qin F, Wang N et al. Dual-function composite optical fiber Fabry–Pérot sensor and its demodulation system for simultaneous high-precision measurement of pressure and temperature. IEEE Trans Instrum Meas 74, 9537007 (2025). DOI: 10.1109/TIM.2025.3615309

57. 57. Tan JH, Qin F, Wang N et al. All-sapphire-based high-temperature pressure sensor system with in situ temperature compensation: innovative cavity design, fabrication, and APSC-FFT algorithm. Microsyst Nanoeng 12, 159 (2026). DOI: 10.1038/s41378-026-01290-5

58. 58. Qin F, Tan JH, Guo JT et al. Fiber-optic Fabry-Pérot interferometric accelerometer with composite cavity and temperature calibration for high-temperature and high-pressure applications. Microsyst Nanoeng 12, 155 (2026). DOI: 10.1038/s41378-026-01250-z

59. 59. Li ZR, Dang WJ, Dan JX et al. High-sensitivity interferometric high-temperature strain sensor based on optical harmonic vernier effect. Opt Fiber Technol 79, 103361 (2023). DOI: 10.1016/j.yofte.2023.103361

60. 60. Zhao XN, Wu XQ, Zuo C et al. Miniatured Fabry-Perot strain probe based on anti-resonant hollow core fiber. Measurement 203, 111775 (2022). DOI: 10.1016/j.measurement.2022.111775

61. 61. Wang X, Jiang JF, Wang S et al. All-silicon dual-cavity fiber-optic pressure sensor with ultralow pressure-temperature cross-sensitivity and wide working temperature range. Photonics Res 9, 521–529 (2021). DOI: 10.1364/PRJ.414121

62. 62. Zhang YT, Jiang Y, Yang SW et al. All-sapphire fiber-optic sensor for the simultaneous measurement of ultra-high temperature and high pressure. Opt Express 32, 14826–14836 (2024). DOI: 10.1364/OE.519656

63. 63. Cui Y, Jiang Y, Zhang YT et al. Sapphire optical fiber high-temperature vibration sensor. Opt Express 30, 1056–1065 (2022). DOI: 10.1364/OE.447449

64. 64. Qiu ZJ, Zhang GQ, Zhang M et al. Research on a signal demodulation algorithm for fiber optic acoustic sensors based on the amplitude ratio of harmonic components. Opt Express 32, 42274–42288 (2024). DOI: 10.1364/OE.537932

65. 65. Yoshino T, Kurosawa K, Itoh K et al. Fiber-optic Fabry-Perot interferometer and its sensor applications. IEEE Trans Microw Theory Tech 30, 1612–1621 (1982). DOI: 10.1109/TMTT.1982.1131298

66. 66. Miao YZ, Xue F, Li MW et al. Spectrum contrast enhancement of fiber Fabry-Perot sensor by coupling efficiency improvement. Opt Commun 573, 131015 (2024). DOI: 10.1016/j.optcom.2024.131015

67. 67. Dandridge A, Tveten A, Giallorenzi T. Homodyne demodulation scheme for fiber optic sensors using phase generated carrier. IEEE J Quantum Electron 18, 1647–1653 (1982). DOI: 10.1109/JQE.1982.1071416

68. 68. Li Q, Shao J, Wu P et al. Research on high-precision PGC demodulation method for Fabry-Perot sensors based on shifted sampling pre-calibration. Sensors 25, 5990 (2025). DOI: 10.3390/s25195990

69. 69. Chen YC, Taylor HF. Multiplexed fiber Fabry–Perot temperature sensor system using white-light interferometry. Opt Lett 27, 903–905 (2002). DOI: 10.1364/OL.27.000903

70. 70. Xu YF, Qi HC, Zhao XY et al. High-speed spectrum demodulation of fiber-optic Fabry–Perot sensor based on scanning laser. Opt Lasers Eng 178, 108192 (2024). DOI: 10.1016/j.optlaseng.2024.108192

71. 71. Li WC, Yuan YG, Yang J et al. Review of optical fiber sensor network technology based on white light interferometry. Photonic Sens 11, 31–44 (2021). DOI: 10.1007/s13320-021-0613-x

72. 72. Pan Z, Liu HJ, Wang LX et al. Draw-tower-grating-based Fabry-Perot interferometers for simultaneous measurements of temperature and salinity. Opt Express 32, 22687–22699 (2024). DOI: 10.1364/OE.523935

73. 73. Zhu ZD, Ba DX, Liu L et al. Temperature-compensated multi-point refractive index sensing based on a cascaded Fabry-Perot cavity and FMCW interferometry. Opt Express 29, 19034–19048 (2021). DOI: 10.1364/OE.424995

74. 74. Chen YF, Wang TH, Yang PC et al. Optical Vernier effect-based demodulation system for fiber-optic sensors: an AWG and a neural network-based perspective. Appl Opt 64, 6077–6084 (2025). DOI: 10.1364/AO.563160

75. 75. Ren SF, Chen SC, Hu YL et al. Rethink training: synthetic data powers efficient fiber-optics sensor demodulation neural networks. IEEE Sens J 24, 31408–31416 (2024). DOI: 10.1109/JSEN.2024.3441239

76. 76. Wang L, Zhou R, Li M et al. Asymmetric reflection spectrum of Fabry-Perot interferometer and the application in pressure sensing. J Lightwave Technol 40, 4011–4019 (2022). DOI: 10.1109/JLT.2022.3153517

77. 77. Zuo Y, Li C, Zhao Y et al. A dual-wavelength demodulation-based sensor for magnetic fields. Sens Actuators A Phys 366, 114959 (2024). DOI: 10.1016/j.sna.2023.114959

78. 78. Li ZX, Wan SP, Yu JS. Research on real-time gear fault detection and classification technology based on EFPI vibration sensor. Opt Laser Technol 192, 113627 (2025). DOI: 10.1016/j.optlastec.2025.113627

79. 79. Ye W, Ren QY, Wang HR et al. High-stability quadrature-point demodulation technique for dynamic interrogation of fiber Fabry–Perot sensors. Appl Opt 65, 1427–1436 (2026). DOI: 10.1364/AO.585838

80. 80. Wang F, Guo XH, Yuan YZ et al. Temperature monitoring system for fiber optic multichannel applications based on double-cone microstructure and ceramic-inserted F-P cavities. Opt Laser Technol 198, 114913 (2026). DOI: 10.1016/j.optlastec.2026.114913

81. 81. Zhang SH, Zhang LQ, He C et al. Improved PGC-Arctan demodulation scheme with immunity to carrier phase delay and modulation depth. Opt Lasers Eng 176, 108072 (2024). DOI: 10.1016/j.optlaseng.2024.108072

82. 82. Liu Y, Qi B, Winder DE. Faraday Michelson interferometers for signal demodulation of fiber-optic sensors. J Lightwave Technol 39, 2552–2558 (2021). DOI: 10.1109/JLT.2021.3053370

83. 83. Zhang WJ, Lu P, Qu ZY et al. Four-wavelength quadrature phase demodulation technique for extrinsic Fabry–Perot interferometric sensors. Opt Lett 47, 2406–2409 (2022). DOI: 10.1364/OL.460132

84. 84. Shi JH, Guang D, Li SL et al. Single-wavelength passive phase-shifted demodulation technique with the dual-cavity and EFA for the interrogation of EFPI diaphragm-based fiber sensor. J Lightwave Technol 40, 222–227 (2022). DOI: 10.1109/JLT.2021.3069033

85. 85. Huang Y, Chen HW, Zhang LN et al. Improved dual-wavelength demodulation technique based on prior information for fiber optic Fabry-Pérot ultrasonic sensors. J Lightwave Technol 42, 6444–6450 (2024). DOI: 10.1109/JLT.2024.3416989

86. 86. Zhang SH, Dong JY, Peng M et al. Fiber-optic extrinsic Fabry–Perot interferometer sensors with quadrature phase three wavelength demodulation. Eur Phys J Plus 139, 1005 (2024). DOI: 10.1140/epjp/s13360-024-05809-5

87. 87. Lee CE, Taylor HF. Fiber-optic Fabry-Perot temperature sensor using a low-coherence light source. J Lightwave Technol 9, 129–134 (1991). DOI: 10.1109/50.64932

88. 88. Guo ZL, Lv WT, Wang W et al. Absolute single cavity length interrogation of fiber-optic compound Fabry–Perot pressure sensors through a white light non-scanning correlation method. Sensors 19, 1628 (2019). DOI: 10.3390/s19071628

89. 89. Tan JH, Wang Q, Wang N et al. Hardware optical correlation method for dynamic detection of micrometer gaps with sub-nanometer resolution. Opt Express 30, 32660–32669 (2022). DOI: 10.1364/OE.464475

90. 90. Kong DY, Song ZY, Wang N et al. Fiber Fabry-Perot demodulation system based on dual Fizeau interferometers. Photonic Sens 13, 230229 (2023). DOI: 10.1007/s13320-022-0670-9

91. 91. Chen HB, Zhang Y, Li Y et al. Squared peak-to-peak algorithm for the spectral interrogation of short-cavity fiber-optic Fabry–Perot sensors. Appl Opt 59, 1198–1205 (2020). DOI: 10.1364/AO.382169

92. 92. Jiang Y. High-resolution interrogation technique for fiber optic extrinsic Fabry-Perot interferometric sensors by the peak-to-peak method. Appl Opt 47, 925–932 (2008). DOI: 10.1364/AO.47.000925

93. 93. Mei JW, Xiao XS, Yang CX. High-resolution and large dynamic range fiber extrinsic Fabry–Perot sensing by multi-extrema-tracing technique. Appl Opt 54, 3677–3681 (2015). DOI: 10.1364/AO.54.003677

94. 94. Hai ZY, Su ZX, Liang R et al. Numerical and experiment analysis of sapphire sandwich-structure Fabry–Perot pressure sensor through fast Fourier transform and mean square error demodulation algorithm. Materials 17, 3649 (2024). DOI: 10.3390/ma17153649

95. 95. Shao ZQ, Wu YL, Sun ZQ et al. Excellent repeatability, all-sapphire Fabry Perot optical pressure sensor based on wet etching and direct bonding for harsh environment applications. Opt Express 29, 19831–19838 (2021). DOI: 10.1364/OE.423753

96. 96. Li CX, Qi HC, Han X et al. Ultrahigh-speed phase demodulation of a Fabry–Perot sensor based on fiber array parallel spectral detection. Opt Lett 49, 714–717 (2024). DOI: 10.1364/OL.511903

97. 97. Shin DW, Matsukuma H, Sato R et al. Improved peak-to-peak method for cavity length measurement of a Fabry-Perot etalon using a mode-locked femtosecond laser. Opt Express 31, 25797–25814 (2023). DOI: 10.1364/OE.493507

98. 98. Zhang T, Han Q, Ma QY et al. Intensity demodulation of optical vernier effect: an intelligent algorithm approach. IEEE Sens J 24, 10084–10090 (2024). DOI: 10.1109/JSEN.2024.3363019

99. 99. Li JW, Gao MQ, Li JL et al. High accuracy sitting posture monitoring enabled by multiplexing fiber optic microsphere array and deep learning. Measurement 268, 120710 (2026). DOI: 10.1016/j.measurement.2026.120710

100. 100. Li YC, Liu B, Xiao S et al. Dual-parametric demodulation of temperature and pressure for EFPI-FBG sensors based on deep learning methods. Opt Lasers Eng 198, 109482 (2026). DOI: 10.1016/j.optlaseng.2025.109482

101. 101. Liu J, Qian J, Zhao QR et al. All-silica optical fiber Fabry-Perot vibration sensor based on MEMS and laser welding for high temperature up to 800 °C. Sens Actuators A Phys 399, 117497 (2026). DOI: 10.1016/j.sna.2026.117497

102. 102. Li YH, Wang J, Sun YN et al. Fiber-optic Fabry-Perot sensor based on spectral quality-enhanced microsphere lens for high-temperature strain measurement. Opt Express 34, 5982–5996 (2026). DOI: 10.1364/OE.582907

103. 103. Zhu YM, Zhou JM, Jin T et al. High-temperature-resistant gas pressure sensor based on an all-silica Fabry-Perot interferometer. Appl Opt 64, 5688–5697 (2025). DOI: 10.1364/AO.565014

104. 104. He X, He ZX, Ran ZL et al. Temperature-insensitive Quasi-distributed Fiber-optic Fabry-Perot high-pressure sensing based on microwave interference system. Opt Laser Technol 161, 109069 (2023). DOI: 10.1016/j.optlastec.2022.109069

105. 105. He X, Ran ZL, Ding ZY et al. Fast-response sensitivity-enhanced all-fiber Fabry-Pérot high-pressure sensor for internal combustion engines. J Lightwave Technol 40, 299–307 (2022). DOI: 10.1109/JLT.2021.3119410

106. 106. Dai GH, Su D, Qiao XG. Sensitivity-enhanced high-pressure sensor based on suspended fiber-optic Fabry-Pérot interferometer. IEEE Trans Instrum Meas 71, 7008507 (2022). DOI: 10.1109/TIM.2022.3217847

107. 107. Zhao SS, Sun QQ, Yang SQ et al. High-resolution silicon-based optical Fabry-Perot sensing system for seawater temperature measurement. J Lightwave Technol 43, 10312–10321 (2025). DOI: 10.1109/JLT.2025.3608331

108. 108. Wang EB, Niu HJ, Li P et al. Optical fiber F-P cavity pH sensor based on polyaniline reaction deposition film layer. Measurement 243, 116454 (2025). DOI: 10.1016/j.measurement.2024.116454

109. 109. Yang HR, Xie S, Xin F et al. Ultra-sensitive F–P humidity sensor based on an open-cavity note-type structure for GIS low-humidity monitoring. IEEE Sens J 26, 13197–13204 (2026). DOI: 10.1109/JSEN.2026.3677260

110. 110. Zhang JW, Wang L, Chen JH et al. An optical measurement method for nanosecond impulse signal monitoring of partial discharge based on the Fabry-Perot interferometer sensing system. Jpn J Appl Phys 61, 102006 (2022). DOI: 10.35848/1347-4065/ac88a9

111. 111. Hossain MM, Pant TK, Vineeth C. Nocturnal thermospheric neutral wind and temperature measurement using a Fabry-Perot interferometer: first results from an equatorial Indian station. Adv Space Res 72, 598–613 (2023). DOI: 10.1016/j.asr.2023.03.040

112. 112. Hwang J, Ku Z, Jeon J et al. Polarization-sensitive and wide incidence angle-insensitive Fabry–Perot optical cavity bounded by two metal grating layers. Sensors 20, 5382 (2020). DOI: 10.3390/s20185382

113. 113. Pang XK, Yao GZ, Wang ZF et al. Sapphire MEMS fiber-optic sensor with a ceramic-housing package for high-temperature vibration. IEEE Photonics Technol Lett 38, 1231–1234 (2026). DOI: 10.1109/LPT.2026.3682756

114. 114. Gao ZX, Du X, Li YY et al. MOF-based fiber-optic iodine vapor vernier sensor with temperature crosstalk reduction by TOC&TEC. Opt Fiber Technol 94, 104381 (2025). DOI: 10.1016/j.yofte.2025.104381

115. 115. Zhang YT, Jiang Y, Cui Y et al. An all-fiber diaphragm-based extrinsic Fabry-Perot sensor for the measurement of pressure at ultra-low temperature. Meas Sci Technol 33, 055117 (2022). DOI: 10.1088/1361-6501/ac51f0

116. 116. Li SL, Chen C, Zhao DL et al. A dynamic spectral demodulation scheme of the fiber optical Fabry-Perot sensor for low-frequency vibration detection. Opt Laser Technol 171, 110436 (2024). DOI: 10.1016/j.optlastec.2023.110436

117. 117. Amoudry L, Wang H, Cassou K et al. Modal instability suppression in a high-average-power and high-finesse Fabry-Perot cavity. Appl Opt 59, 116–121 (2020). DOI: 10.1364/AO.59.000116

118. 118. Xie BW, Zhang WJ, Zhao JM et al. VO₂-based superposed Fabry-Perot multilayer film with a highly enhanced infrared emittance and emittance tunability for spacecraft thermal control. Opt Express 30, 34314–34327 (2022). DOI: 10.1364/OE.464266

119. 119. Allam AM, Mahmoud AM, Boltia SA et al. Photonic sensor for selective Ticagrelor detection based on porous silicon Fabry-Pérot interferometer modified with molecularly imprinted co-polymer. Microchem J 214, 114007 (2025). DOI: 10.1016/j.microc.2025.114007

120. 120. Chen H, Buric M, Ohodnicki PR et al. Review and perspective: sapphire optical fiber cladding development for harsh environment sensing. Appl Phys Rev 5, 011102 (2018). DOI: 10.1063/1.5010184

121. 121. Ferreira MFS, Rehan M, Mishra V et al. Roadmap on specialty optical fibers. J Phys Photonics 7, 012501 (2025). DOI: 10.1088/2515-7647/ad6b19

122. 122. Mondal A, El Aadad H, El Hamzaoui H et al. Highly stable 3D printed microsensors at optical fiber tips. Sens Actuators A Phys 390, 116620 (2025). DOI: 10.1016/j.sna.2025.116620

123. 123. Liu J, Zhang L, Wang ZY et al. Extreme dual-parameter optical fiber sensor composed of MgO Fabry-Perot composite cavities for simultaneous measurement of temperature and pressure. Appl Sci 15, 8891 (2025). DOI: 10.3390/app15168891

124. 124. Wang SX, Lei ZZ, Jiang YG. Research progress on ceramic-based high-temperature-resistant fiber-optic sensor technology. J Funct Mater Dev 32, 022102 (2026). DOI: 10.3724/jfmd.2507042

125. 125. Chen XY, Kang X, Gan Y et al. High-resolution, broadband on-chip spectrometer with tunable multi-mode waveguide Bragg gratings. Opt Express 33, 35385–35399 (2025). DOI: 10.1364/OE.570103

126. 126. Chen XY, Gan XT, Zhu Y et al. On-chip micro-ring resonator array spectrum detection system based on convex optimization algorithm. Nanophotonics 12, 715–724 (2023). DOI: 10.1515/nanoph-2022-0672

127. 127. Zhang ZM, Xiao SM, Song QH et al. Scalable on-chip diffractive speckle spectrometer with high spectral channel density. Light Sci Appl 14, 130 (2025). DOI: 10.1038/s41377-025-01797-y

128. 128. 128 Shojaei Khatouni S, Zakowski S, Hosseini Vedad R et al. Discriminating temperature, strain, and bending effects in fiber Bragg grating Fabry-Perot interferometer sensors using traditional machine learning techniques. V1. TechRxiv (2023-04-19)[2026-05-28]. https://doi.org/10.36227/techrxiv.22644031.v1.

129. 129. Deng CF, Yu MX, Zhu LQ et al. A deep learning algorithm ADPNet for strain and temperature decoupling of fiber Bragg gratings. Opt Fiber Technol 79, 103356 (2023). DOI: 10.1016/j.yofte.2023.103356

130. 130. Cao LD, Abedin S, Cui GQ et al. Artificial intelligence and machine learning in optical fiber sensors: a review. Sensors 25, 7442 (2025). DOI: 10.3390/s25247442

131. 131. 131 Ren SF, Chen SC, Shi H et al. A generalizable wavelength demodulation method for fiber-optic Fabry-Pérot pressure sensors via federated learning. In <italic>Proceedings of the 23rd International Conference on Optical Communications and Networks</italic> 1–3 (IEEE, Zhangjiajie, 2025). https://doi.org/10.1109/ICOCN67308.2025.11145411.

132. 132. Wei CH, Liu Q, Wang YP et al. Accurate strain sensing with fiber-optic Fabry-Perot sensors based on CNN-LSTM model. IEEE Sens J 24, 17725–17732 (2024). DOI: 10.1109/JSEN.2024.3386724

133. 133. Mo CY, Yang YQ, Mu XG et al. Sensitive detection of multi-point temperature based on FMCW interferometry and DSP algorithm. Nanomaterials 15, 1545 (2025). DOI: 10.3390/nano15201545

134. 134. Malinka F, Vachek O, Krizan D et al. Advanced intensity-modulated fiber sensors for scalable sensing. iScience 28, 113670 (2025). DOI: 10.1016/j.isci.2025.113670

135. 135. 135 Micron Optics. ENLIGHT sensing analysis software. Available online: https://lunainc.com/sites/default/files/assets/files/resource-library/Micron-Optics-ENLIGHT-0902.b_0902.3.pdf (accessed on 2026-05-28).

136. 136. 136 Micron Optics. sm125 Optical sensing interrogator. Available online: https://pdf.directindustry.com/pdf/micron-optics/sm125/101517-610145.html (accessed on 2026-05-28).

137. 137. Yang B, Zhao QC, Che CR et al. High-precision optical fiber temperature and pressure sensor and its application in production wells. Laser Optoelectron Prog 59, 1706006 (2022). DOI: 10.3788/LOP202259.1706006

138. 138. Liu Y, Wen ZX, Yang CX et al. A high-precision and high-frequency dynamic strain measurement system for aeroengine. Meas Control Technol 43, 63–69 (2024). DOI: 10.19708/j.ckjs.2023.11.261

139. 139. Zhuang Y, Zou J, Zhang JQ et al. On-chip sub-picometer continuous wavelength fiber-Bragg-grating interrogator. Photonic Sens 14, 240126 (2024). DOI: 10.1007/s13320-023-0694-9

140. 140. Marin YE, Nannipieri T, Oton CJ et al. Integrated FBG sensors interrogation using active phase demodulation on a silicon photonic platform. J Lightwave Technol 35, 3374–3379 (2017). DOI: 10.1109/JLT.2016.2598395

141. 141. 141 VIAVI Solutions Inc. VIAVI launches industry-leading true-phase DAS fiber sensing interrogator with AI/ML at the edge to enable real-time infrastructure monitoring. Available online: https://newswire.telecomramblings.com/2026/03/viavi-launches-industry-leading-true-phase-das-fiber-sensing-interrogator-with-ai-ml-at-the-edge-to-enable-real-time-infrastructure-monitoring/ (accessed on 2026-05-28).

2026-0004_online.pdf

Abstract

Integrating intraoperative imaging guidance with in situ acquisition of tissue functional information is of great importance for precision interventional diagnosis and therapy of deep-seated lesions, yet achieving both within a single compact fiber probe without compromising either modality remains challenging. Here we report a dual-modal all-fiber needle probe integrating forward-viewing optical coherence tomography (OCT) with in situ fiber Bragg grating (FBG) temperature sensing in a single-mode fiber-FBG-multimode fiber (MMF) architecture. Multimode interference in the MMF section generates a quasi-Bessel beam with a lateral resolution of 5 μm and a depth of focus of 1.25 mm, representing a 7.9-fold improvement over conventional Gaussian-beam probes. The FBG centered at 1550 nm provides photobleaching-free temperature sensing at 10.68 pm/°C without perturbing the OCT channel. The probe is validated across multiple biological tissue specimens and applied to a murine breast tumor model undergoing in vivo photothermal therapy (PTT). In situ FBG-recorded intratumoral temperatures exceed surface thermal camera readings by approximately 1 °C at peak heating, underscoring the indispensable role of co-localized thermometry for accurate thermal dose control. In addition, periodic artifacts observed in M-mode OCT images may offer a potential route for respiration monitoring, requiring no additional sensor hardware. This dual-modal probe establishes a versatile platform for image-guided, thermometry-monitored interstitial cancer therapy.

Keywords

optical coherence tomography; optical fiber sensor; photothermal therapy

1 Introduction

Precision interventional diagnosis and therapy, increasingly enabled by the optical fiber needle probes advanced to lesions deep within the body, has become central to the minimally invasive treatment of tumors and other deep-seated diseases^1,2. Compared with open surgery, such fiber-based intervention with imaging guidance confines diagnosis and treatment to a controlled, targeted region, offering markedly lower procedural risk and greater organ preservation³, and it underpins a broad range of clinical ablative therapies, including radiofrequency⁴ and microwave⁵ thermal ablation, cryoablation⁶, and laser-mediated hyperthermia such as photothermal therapy (PTT)^7,8. The success of these procedures hinges on accurate intraoperative imaging guidance for probe placement, together with the simultaneous acquisition of tissue functional information, such as in situ temperature, for controlled and safe treatment. However, conventional clinical modalities such as computed tomography⁹ and ultrasound¹⁰ provide semi-blind intraoperative guidance, offering only two-dimensional representations of inherently three-dimensional anatomy with limited resolution and lacking real-time, depth-resolved structural feedback at the advancing probe tip. Integrating both structural imaging guidance and in situ functional sensing within a single compact fiber needle probe therefore remains a major unmet need.

Optical coherence tomography (OCT) directly addresses this need for intraoperative imaging guidance, providing label-free, micrometer-scale and depth-resolved structural imaging that is intrinsically compatible with optical fiber delivery and well suited to forward-viewing needle probes¹¹. However, these OCT fiber imaging probes are limited to structural imaging and do not provide complementary functional information which is essential for a more comprehensive understanding of complex biological processes^12–14. Li et al. proposed and demonstrated a single-fiber needle probe that combines OCT imaging with in situ temperature sensing, where a rare-earth-doped tellurite-glass lens integrated onto a double-clad fiber (DCF) enables depth-resolved imaging and fluorescence-based thermometry¹⁵. However, fluorescence-based thermometry may raise safety concerns, as the required 980 nm excitation is strongly absorbed by water and heats the surrounding tissue¹⁶. The sensing channel is not completely isolated from the OCT channel, and thus intensity fluctuations in the OCT source shift the measured fluorescence ratio and may introduce errors into the temperature readout¹⁵. More importantly, existing OCT probes primarily employ Gaussian-beam imaging, which is fundamentally constrained by the trade-off between lateral resolution and depth of focus (DOF), preventing high-resolution imaging over an extended depth range. Hence, achieving high-resolution, large DOF imaging guidance and in situ temperature sensing within a single fiber needle probe without compromising either modality remains challenging.

Fiber Bragg grating (FBG) temperature sensors have been widely deployed in biomedical applications^17,18, and their all-fiber architecture is well suited to multimodal integration within a single OCT probe. In this work, we propose and experimentally demonstrate an integrated all-fiber OCT–FBG sensing probe that combines intraoperative imaging guidance with in situ temperature monitoring while minimizing interference between the imaging and sensing channels. The probe is built on an single-mode fiber (SMF)-FBG-multimode fiber (MMF) architecture, exploiting multimode interference within a short MMF section to generate a quasi-Bessel beam with a lateral resolution of 5 μm, a DOF of 1.25 mm, and a 7.9-fold improvement over conventional Gaussian-beam probes. The FBG operating at 1550 nm provides a temperature sensitivity of 10.68 pm/°C without perturbing the OCT channel. The probe is validated across multiple biological specimens and in a murine breast tumor model, where the in situ FBG temperature exceeds the surface thermal camera reading by ~1 °C at peak heating, underscoring the importance of co-localized intratumoral thermometry. The periodic M-mode OCT artifacts may be leveraged as a potential signal for respiratory monitoring during therapy. This dual-modal probe establishes a versatile platform for image-guided, thermometry-monitored interstitial PTT and broader minimally invasive interventional applications.

2 Results

2.1 Dual-modal all-fiber OCT probe

Conventional OCT fiber probes rely on a SMF-no core fiber (NCF)-graded-index (GRIN) fiber structure to focus a Gaussian beam, in which the resolution and the DOF are intrinsically limited by the GRIN fiber¹⁹. Here, instead, we generate a quasi-Bessel beam by exploiting the coherent superposition of multiple high-order modes excited in a MMF²⁰, yielding a simpler and more compact all-fiber structure with an extended DOF. This enables the realization of an MMF-based all-fiber probe architecture that is compact and readily compatible with additional fiber-optic sensing components, facilitating seamless integration with the FBG sensors. Hence, as depicted in Fig. 1(a), an all-fiber probe designed with an SMF-FBG-MMF structure is proposed for simultaneous OCT imaging and temperature sensing. The probe incorporates an 8 mm UV-laser-inscribed FBG at the probe tip, with a Bragg reflection wavelength centered at 1550 nm. This wavelength lies well outside the OCT operating band at ~1310 nm, ensuring the FBG functions exclusively as a temperature sensing element without perturbing the OCT spectral response. Figure 1(d) shows the fabricated all-fiber probe encapsulated within a medical injection needle for interventional imaging (details of the transmission mechanism and fabrication of the all-fiber OCT probe are provided in the Methods section).

Figure 1. The dual-modal all-fiber OCT probe. (a) Schematic of the all-fiber OCT probe with a SMF-FBG-MMF structure. (b) Simulated xz beam profile and FWHM variation of the all-fiber probe with MMF core diameters of 50 and 105 μm at an incident wavelength of 1310 nm. The two probes are designed with the same MMF length of 2 mm. (c) Measured beam profile of the fabricated all-fiber probe with MMF lengths of 1.5, 2 and 2.5 mm with core diameters of 105 μm. (d) The encapsulated OCT needle probe. The subgraph shows the micro image of the fiber probe with the FBG illuminated by a 650 nm laser. Scale bar in subgraph of (d) represents 100 μm. (e) Measured lateral intensity distribution of the 2 mm MMF all-fiber probe at the focal plane. (f) Measured beam FWHM variation of the 2 mm MMF all-fiber probe as a function of propagation distance. The subgraphs show the lateral beam spot at the propagation distance of 0, 0.5 and 1.5 mm. (g) Measured central intensity variation of the 2 mm MMF all-fiber probe along propagation distance. The blue dotted line marks the beam intensity level down to 1/e².

The beam profiles of the OCT probe with different MMF core diameters of 50 and 105 μm were simulated with a beam propagation method (BPM)²¹, and variations of full-width at half-maximum (FWHM) beam diameter are presented in Fig. 1(b), respectively. As the core diameter increases from 50 μm to 105 μm, the output beam more closely approximates a quasi-Bessel beam, with the divergence angle θ reduced from 3.8° to 1.0°. Figure 1(c) presents the measured beam profiles of the all-fiber probe with the MMF lengths of 1.5, 2, and 2.5 mm with core diameters of 105 μm, using a customized beam profile analyzing system²². The oscillatory variations in central intensity and side-lobe variations along the propagation distance are consistent with the theoretical expectations of multimode interference, confirming that the output beam approximates a quasi-Bessel beam. The results further demonstrate that both the lateral resolution and DOF are tunable through adjustment of the MMF length.

The all-fiber probe with an MMF length of 2 mm and the core diameter of 105 μm yields the most pronounced DOF extension and was accordingly selected for all subsequent imaging and PTT experiments. Figure 1(e) plots the normalized intensity distribution along the x- and y-axes at the focal plane, with measured beam diameters of 4.8 μm along both axes. The side-lobe intensities are 18.2% and 22.0% of the central peak, significantly lower than the central spot intensity, ensuring that no substantial imaging artifacts would be introduced. As shown in Fig. 1(f), the all-fiber probe generates a quasi-Bessel beam with an extended DOF (defined as the axial range over which the beam FWHM doubles from its minimum) of 1.25 mm, achieving approximately a 7.9-fold enhancement compared to a conventional GRIN fiber probe¹⁹ based on Gaussian beams. Furthermore, the effective axial depth (defined as the region where the central intensity remains above 1/e²) is 1.5 mm, ensuring adequate energy uniformity and continuity over an extended range despite the oscillations in the central intensity, as illustrated in Fig. 1(g).

2.2 Dual-modal imaging and temperature sensing system

To simultaneously achieve endoscopic imaging and temperature monitoring, a dual-modal system was developed based on a customized swept-source OCT (SS-OCT) system capable of continuous M-mode OCT imaging¹⁵, as illustrated in Fig. 2(a) (details can be found in the Methods section). An additional sensing arm was incorporated in the system, including an amplified spontaneous emission (ASE) source, a wavelength division multiplexer (WDM, WD6513A, Thorlabs, USA), a 99:1 fiber coupler (TW1300R1A1, Thorlabs, USA) and an optical spectrum analyzer (OSA) for FBG interrogation. Figure 2(b) illustrates the optical spectra of port 1, 2 and 3. The spectral response within the 1260–1360 nm range originates from the swept source, while the 1527–1565 nm range is attributed to the ASE source. The spectral power of port 2 is comparable to that of port 3. This balance ensures that interferometric demodulation is not compromised upon entering the balanced photodetector (BPD). A high-contrast reflection peak with a bandwidth of 0.3 nm and an intensity of 22 dB is excited at 1550 nm by the FBG.

Figure 2. Characterization of the OCT imaging and temperature sensing system. (a) Schematic of the SS-OCT based imaging and sensing system. ASE source: Amplified spontaneous emission source. WDM: Wavelength division multiplexer. OSA: Optical spectrum analyzer. (b) Optical spectra of port 1, 2 and 3. (c) Experimental setup of M-mode OCT imaging of coverslips and water temperature sensing. (d) Spectral responses of the FBG sensor before encapsulation, with the inset of wavelength shift as the temperature increases. (e) Spectral responses of the FBG sensor after encapsulation, with the inset of wavelength shift as the temperature increases. (f) M-mode OCT image of coverslips. Scale bar in f represents a time period of 1 s.

As shown in Fig. 2(c), imaging and temperature sensing experiments were conducted by positioning the OCT probe vertically above a stack of coverslips submerged in a water tank. For temperature characterization, the probe was immersed in water, and the water temperature was varied from 20 to 70 °C using a cooling plate (SCP-85, As ONE, Japan). The water temperature was simultaneously monitored using a thermocouple temperature sensor (175 H1, Testo, Germany) as a reference, while the temperature sensitivity of the FBG was calculated from the spectral shift of the reflection peak. Figure 2(d) and 2(e) show the spectral responses of a bare fiber probe and an encapsulated OCT needle probe, respectively. In both cases, the FBG reflection peak exhibited a redshift with increasing temperature. The bare fiber probe and encapsulated OCT needle probe exhibited temperature sensitivities of 10.68 and 10.61 pm/°C, respectively, which are consistent with those of standard FBG temperature sensors²³. The needle encapsulation structure has a negligible effect on the temperature sensitivity of the FBG sensor. The M-mode OCT image in Fig. 2(f) demonstrates the dynamic transition of the fiber probe as it moves from air into water and subsequently approaches the coverslips. A total of about 26 layers of coverslips are distinguishable in Fig. 2(f), demonstrating an effective imaging depth of circa 1.3 mm in such low-scattering media. These results demonstrate that the proposed probe is a compact dual-modal platform suitable for real-time forward-viewing imaging guidance and in situ temperature monitoring during precision interventions.

2.3 Imaging characterization

Further imaging experiments were performed to validate the potential of the quasi-Bessel beam generated by the all-fiber probe in extended-DOF imaging. Figure 3(a) and 3(b) present the 3D OCT volumetric images of transparent adhesive tape reconstructed using the proposed probe and a standard GRIN fiber probe¹⁹, respectively. The GRIN probe, serving as a counterpart, was fabricated by sequentially splicing an optimized length of NCF (CL 0/125-0/250, YOFC, China) and GRIN fiber (GI 100/125-29/250, YOFC, China) to an SMF, providing a resolution of 5 μm and a DOF of 140 μm. Combined with the 2D cross-sectional images in Fig. 3(a) and 3(b), it is evident that the proposed all-fiber probe can recover deeper structural details, clearly resolving approximately 15 layers of tape (corresponding to an imaging depth of 1 mm). In contrast, the image quality of the GRIN fiber probe begins to degrade after about 10 layers, where the layered structures become discrete and blurred. This demonstrates that the excited quasi-Bessel beam maintains a robust energy distribution over an extended range to excite backscattered light, characterized by both long DOF and high resolution. To evaluate the feasibility of the probe for imaging within high-scattering biological environments, we performed OCT imaging on porcine intestinal mucosa and mouse brain sample, as presented in Fig. 3(c) and 3(d), respectively. As shown in Fig. 3(c), the all-fiber probe clearly resolves the intricate layer structure of the porcine intestinal wall, including the muscularis propria (MP), submucosa (SM), muscularis mucosae (MM), lamina propria (LP), and the epithelium (EP)²⁴. Additionally, the micro vessel embedded within the mucosal layer is also identifiable. Figure 3(d) displays the 3D OCT reconstruction of a mouse brain sample. The fissure between the cerebrum and the cerebellum is clearly distinguishable, as shown in the top-view image on the right. Furthermore, the inset in the upper-right corner of Fig. 3(d) presents an OCT cross-sectional image of the brain, where the cerebral cortex and the hippocampus can be structurally distinguished. These results confirm that the all-fiber probe maintains its extended DOF imaging characteristics in biological tissue, which is critical for visualizing subsurface structures and improving lesion identification accuracy in clinical diagnostics. Building on these results, ex vivo OCT imaging was performed on normal murine abdominal skin and murine breast tumor to establish tissue-specific structural references for real-time identification of the probe insertion state during in vivo PTT. As shown in Fig. 3(e), the cross-sectional OCT image of normal abdominal skin clearly resolves three distinct layers, corresponding to the epidermis (ED), dermis (DE), and hypodermis (HD). In contrast, the OCT image of breast tumor tissue (Fig. 3(f)) resolves only two distinct regions, including a highly scattering viable tumor parenchyma (VTP) near the surface and a low-scattering necrotic core (NC) in the deeper region. The dense structural characteristics of the VTP result in strong optical attenuation, blocking the penetration of the OCT beam and thereby limiting the imaging depth. The distinct structural contrast observed in Fig. 3(e) and 3(f) can thus assist in determining the insertion state and penetration depth during interventional PTT.

Figure 3. OCT imaging characterization of the proposed all-fiber OCT probe. 3D OCT reconstruction of translucent tape using (a) the proposed all-fiber probe and (b) the traditional GRIN fiber probe. The insets show the cross-section images of the translucent tape. (c) 3D reconstruction of porcine intestinal mucosa, with insets of the physical image, cross-section and top-view OCT image. (d) 3D reconstruction of mouse brain samples, with insets of the physical image, cross-section and top-view OCT image. (e) 3D reconstruction of murine abdomen epithelial tissue, with inset of the physical image and cross-section OCT image. (f) 3D reconstruction of murine breast tumor, with inset of the physical image and cross-section OCT image. ED: Epidermis. DE: Dermis. HD: Hypodermis. VTP: Viable tumor parenchyma. NC: Necrotic core. Scale bars are 0.5 mm.

2.4 OCT imaging guidance and in situ temperature sensing in tumor and PTT procedure

The proposed dual-modal all-fiber probe was employed for imaging guidance in an in vivo PTT targeting the murine abdominal breast tumor (Animal experiment protocol A202400452, approved by the Animal Ethical and Welfare Committee of Shenzhen University, China), enabling real-time localization of the needle insertion site and simultaneous intratumoral temperature monitoring during the therapeutic procedure. Prior to PTT, a subcutaneous breast tumor model was established in mice, and a biocompatible phototheranostic agent²⁵ was injected intravenously into mice and allowed to accumulate in the tumor via systemic circulation, thereby enhancing the photothermal therapeutic efficacy (details can be seen in the Methods section). A modified phototherapy fiber probe²⁵ for light irradiation was prepared to activate the photothermal agent, and the local temperature was monitored by a thermal camera (A300 IR Temperature Sensor, FLIR, USA).

The entire treatment procedure comprised probe insertion, laser irradiation, cooling and probe withdrawal upon completion. The laboratory mice were anesthetized with isoflurane and secured on the experimental stage for the procedure, as shown in Fig. 4(a). The dual-modal probe and the phototherapy fiber probe were sequentially inserted into the tumor, after which 808 nm laser irradiation was delivered through the phototherapy fiber probe to perform interstitial photothermal treatment. Thermal images acquired before and during laser irradiation are shown in Fig. 4(b) and 4(c), revealing an initial mouse body temperature of 31 °C that then rose to 51 °C within the tumor during irradiation. Figure 4(d) presents the M-mode OCT image acquired during probe insertion through the abdominal skin, where the previously identified ED, DE, and low-scattering HD layers are clearly resolved, along with an additional layered structure tentatively identified as the panniculus carnosus (PC). Figure 4(e) shows the OCT image acquired within the tumor, exhibiting layered features consistent with the previously identified high-scattering VTP and low-scattering NC regions. These results demonstrate that the forward-viewing imaging capability of the dual-modal probe facilitates discrimination of intratumoral structures and improves the accuracy of probe placement.

Figure 4. OCT imaging and temperature sensing during photothermal therapy. (a) Photothermal therapy on a mouse breast tumor model. (b) Thermal image before laser activation. (c) Thermal image during laser irradiation. (d) M-mode OCT image during insertion into epidermis. (e) M-mode OCT image during insertion into tumor. (f) Spectral response of the FBG peak during laser activation. (g) Spectral response of the FBG sensor during laser close off. (h) Temperature variation over time demodulated by the dual-modal probe. (i) Temperature variation over time recorded by thermal camera. (j) Frequency profile over time (min. 3–19) of the periodic artifacts extracted from the M-mode OCT images. The three subgraphs demonstrate the observed time periods of 1, 0.67 and 0.5 s from the M-mode OCT images. ED: Epidermis. DE: Dermis. HD: Hypodermis. PC: Panniculus carnosus. VTP: Viable tumor parenchyma. NC: Necrotic core. All scale bars represent time periods of 0.5 s.

Figure 4(f) and 4(g) present the spectral shift of the grating during the laser irradiation and cooling stages. Figure 4(h) and 4(i) present the temporal evolution of intratumoral temperature recorded by the dual-modal probe and the thermal camera, respectively. The complete temperature profile can be divided into five distinct stages: (I) pre-insertion, (II) probe insertion into the tumor (mouse body temperature: 31.69 °C), (III) laser irradiation-induced heating to a peak temperature of 54.81 °C, (IV) post-irradiation cooling back to body temperature upon laser deactivation, and (V) probe withdrawal. The body temperatures measured by the dual-modal probe (31.69 °C) and thermal camera (31.61 °C) are in close agreement. However, at the peak of laser-irradiation heating, the FBG sensor records a markedly higher intratumoral temperature (54.81 °C) compared to the surface thermal imaging reading (53.69 °C). Although the ~1 °C discrepancy observed here reflects the superficial location of the subcutaneous tumor model, this divergence is expected to increase substantially for deeper-seated visceral tumors, where greater overlying tissue attenuation and vascular heat dissipation render surface thermal imaging an increasingly unreliable surrogate for local intratumoral temperature²⁶. These results underscore the critical importance of in situ temperature monitoring via an integrated optical fiber sensor in interstitial laser irradiation scenarios, which enables more precise thermal dose control and reduces the risk of both tumor under-treatment and collateral thermal damage to surrounding tissue.

After the probe was fully inserted and secured, and laser irradiation was initiated, periodic vertical stripe artifacts appeared in the M-mode OCT image, as shown on the right side of Fig. 4(e). During treatment, the probe and tumor remained relatively stationary, and no external mechanical perturbation was intentionally introduced. The periodic signals may therefore reflect endogenous motion, including respiration-related or cardiac motion. For these M-mode OCT images acquired from the experiment, a custom image-processing algorithm was applied to identify the periodic vertical artifacts, which were subsequently converted into time-varying frequency information via discrete Fourier transform, yielding the time-resolved frequency profile shown in Fig. 4(j). Polynomial fitting revealed that the frequency remained below 1 Hz during the initial five minutes under deep anesthesia. Upon manual reduction of the anesthetic dosage, the frequency gradually increased to ~2 Hz, consistent with the normal anesthetized state. These periodic artifacts are possibly attributed to respiration-driven abdominal motion, in which diaphragmatic contraction induces cyclic displacement of the abdominal wall. The observed frequency range of 0.5–2 Hz corresponds well to the physiological respiratory rate of mice under varying depths of isoflurane anesthesia, where deeper anesthesia suppresses breathing to ~30 breaths/min (~0.5 Hz) and lighter anesthesia allows rates approaching 120 breaths/min (~2 Hz)^27–29.

To further verify whether the periodic artifacts could be induced by mechanical disturbances at comparable frequencies, a validation experiment was conducted as illustrated in Fig. 5(a). In this configuration, the OCT needle probe was directly inserted into a dissected ex vivo mouse brain specimen for M-mode imaging. A glass slide fixed at the base was mechanically coupled to a metal diaphragm, which was driven to oscillate at preset frequencies by square wave signals output from a function/arbitrary waveform generator (SDG 2122X, SIGLENT, China), thereby simulating the effect of mechanical disturbances on the M-mode OCT imaging. Three representative periods of 0.88, 0.67 and 0.39 s were extracted from Fig. 5(c, d) and 5(e), and square wave signals at the corresponding frequencies of 1.13, 1.46, and 2.56 Hz were programmed into the waveform generator to produce periodic mechanical displacement, as shown in Fig. 5(b). Evidently, the periodic artifacts with periods of 0.88 s, 0.67 s, and 0.39 s in the M-mode OCT images closely match the image characteristics observed under vibration interference driven by the corresponding square wave signals, as shown in Fig. 5(f, g) and 5(h). These results demonstrate the promising potential of the dual-modal probe for respiration-related motion monitoring during photothermal therapy.

Figure 5. Imaging validation of the periodic artifacts. (a) Experimental setup for the vibration validation experiment. (b) Programmed square wave signals for vibration excitation with different frequencies of 1.13, 1.46 and 2.56 Hz. The square wave signals were configured with an amplitude of 1 V and a pulse width of 0.2 s. M-mode OCT image segments extracted from the photothermal experiment, exhibiting periodic striations with temporal periods of (c) 0.88 s, (d) 0.67 s and (e) 0.39 s, and M-mode OCT images of the ex vivo mouse brain acquired at the corresponding applied vibration frequencies of (f) 1.13 Hz, (g) 1.46 Hz and (h) 2.56 Hz. Scale bars represent time segments of 5 s.

3 Discussion

The dual-modal all-fiber probe proposed in this work integrates forward-viewing OCT imaging with in situ FBG temperature sensing within a single SMF-FBG-MMF architecture, offering simultaneous structural guidance and thermometry for interstitial PTT. The probe is fabricated entirely by standard fusion splicing without chemical etching^30–33 and precise polishing^34,35, and the quasi-Bessel beam characteristics are tunable by adjusting the MMF length, rendering the fabrication process straightforward, reproducible, and adaptable to specific application requirements. Compared with existing all-fiber OCT probes exploiting multimode interference for DOF extension through stepwise core-diameter transitions^36,37, optimized multimode fiber coupling³⁸ and fiber-based spatial filtering^39,40, the SMF-FBG-MMF probe achieves a substantially greater 7.9-fold DOF improvement through quasi-Bessel beam generation at the cleaved MMF tip. The integrated FBG further provides co-located real-time intratumoral temperature monitoring during PTT, with a dual-modal capability that is absent from all prior imaging-only all-fiber probes. The monolithic all-fiber structure is inherently compact and mechanically robust, enabling seamless encapsulation within a hypodermic needle for minimally invasive deployment. The extended-DOF quasi-Bessel beam maintains high lateral resolution and contrast in highly scattering biological environments including intestinal mucosa and brain tissue. During interstitial PTT, forward-viewing M-mode OCT provides real-time visualization of tissue layer transitions during probe insertion, resolving normal skin layers and the panniculus carnosus during transabdominal access and identifying the characteristic high-scattering VTP and low-scattering NC regions upon tumor entry, thereby providing direct structural feedback for accurate probe placement and intratumoral depth localization that addresses the blind-insertion limitation inherent to existing FBG-only thermometry probes.

The in vivo PTT experiment reveals that the FBG-recorded intratumoral temperature exceeds the surface thermal camera reading by approximately 1 °C at peak heating. Although modest in absolute magnitude, this discrepancy carries significant implications for thermal dose control. In the framework of cumulative equivalent minutes at 43 °C, a standard thermal dose metric in thermal ablation therapies⁴¹, the effective dose rate scales exponentially with temperature. Near 54 °C, a 1 °C difference corresponds to approximately a two-fold change in the instantaneous thermal dose rate, meaning that even a small systematic underestimation of intratumoral temperature by surface imaging can lead to substantial under-dosing of the target. Surface-based estimation is further compromised by signal attenuation through overlying tissue layers, vascular heat-sink effects⁴², and tumor microenvironment heterogeneity. The co-localized in situ FBG measurement provides local intratumoral temperature monitoring close to the irradiation site, enabling more accurate thermal dose control and reducing the risk of undertreating the tumor or inducing collateral thermal damage to surrounding healthy tissue.

M-mode OCT imaging during in vivo PTT revealed periodic striations with a frequency ranging from below 1 Hz under deep anesthesia to approximately 2 Hz upon reduction of anesthetic dosage. Validation experiments using controlled mechanical vibration at corresponding frequencies confirmed that periodic mechanical displacement at the probe tip produces stripe patterns closely matching those observed in vivo, supporting the interpretation that these artifacts may arise from respiration-driven abdominal wall displacement⁴³. Although the low-frequency behavior and its modulation by anesthetic depth are more consistent with respiration-related motion than with primary cardiac pulsation, contributions from cardiac pulsation and other endogenous mechanical disturbances, including respiration-coupled vascular or hemodynamic fluctuations, cannot be fully excluded. Direct co-registration with a dedicated respiratory monitor will therefore be required to establish a definitive physiological correspondence. These findings suggest that the interstitial OCT probe may provide a hardware-free indicator of respiration-related motion during PTT⁴⁴, with potential relevance to anesthetic-depth assessment.

Compared with previously reported single-fiber probes for imaging and sensing, the proposed system presents notable practical advantages. DCF-based probes combining OCT with fluorescence thermometry¹⁵ or pH sensing^12,14 require specialized double-clad fiber couplers and multi-band illumination spanning visible to near-infrared wavelengths. The required excitation wavelength also raises tissue-heating concerns, and incomplete channel isolation allows source fluctuations to introduce errors into the readout. Moreover, they remain susceptible to photobleaching-induced drift in long-term monitoring. By contrast, the proposed FBG-based sensing module requires only the addition of a WDM coupler and an ASE source to the existing OCT infrastructure, while operating entirely within the standard SMF platform. The FBG provides photobleaching-free, wavelength-encoded, millisecond-response temperature readout that is inherently stable over extended monitoring periods. A current limitation of the system is that the single-point FBG provides temperature information only at a fixed location along the probe axis. Multiplexed chirped FBG⁴⁵ or phase-shifted FBG configurations interrogated by OFDR²³ would enable distributed temperature profiling, more comprehensively characterizing the thermal dose distribution within heterogeneous tumors. Additionally, the present M-mode configuration acquires one-dimensional structural information. Extension to three-dimensional imaging via rotational or lateral scanning would further enhance spatial guidance precision.

The dual-modal capability demonstrated here, combining image-guided probe placement with real-time in situ thermometry in a minimally invasive all-fiber package, establishes a promising platform not only for interstitial PTT of breast cancer but also for other thermally ablative interventions, including laser ablation of liver, thyroid, and prostate tumors, where accurate probe positioning and intratumoral temperature control are equally critical. Beyond manual operation, the monolithic all-fiber design and its co-localized structural and thermal feedback make the probe well suited to robot-assisted image-guided intervention platforms⁴⁶, where this real-time feedback could enable automated and high-precision needle placement and closed-loop control of the ablation process. With further refinement of probe dimensions, formal biocompatibility validation, and integration of closed-loop thermal dosimetry algorithms, this system has the potential to advance toward clinical translation as an integrated tool for image-guided minimally invasive cancer therapy.

4 Methods

4.1 Propagation mechanism of the all-fiber OCT probe

The propagation mechanism of the proposed OCT probe can be explained with the self-imaging effect of multimode interference theory^47,48. In the MMF core, both the fundamental LP_0,1 mode and higher-order LP_0,n modes are simultaneously excited and guided. Each mode can be represented as an apertured Bessel function {J}_{0}({k}_{z,n}r)\({J}_{0}({k}_{z,n}r)\) with a characteristic transverse wave vector {k}_{z,n}={({n_{f}^{2}}{{k}^{2}}-{\beta _{n}^{2}})}^{1/2}\({k}_{z,n}={({n_{f}^{2}}{{k}^{2}}-{\beta _{n}^{2}})}^{1/2}\), where k = 2π/λ is the free-space wave number, n_f is the refractive index of the MMF core, and r is the radial coordinate in the transverse plane bounded by the MMF core radius. Since each LP_0,n mode travels independently along the fiber with a distinct propagation constant β_n, the total field at the MMF output facet EMMF is obtained as the coherent superposition of all Bessel-like fields²⁰:

{E}_{MMF}\left(r,{L}_{MMF}\right)=\sum \limits_{n=1}^{N}{C}_{n,MMF}{J}_{0}\left({k}_{z,n}r\right)\,\,{e}^{i{{\beta }_{n}}{{L}_{MMF}}}\,,\[{E}_{MMF}\left(r,{L}_{MMF}\right)=\sum \limits_{n=1}^{N}{C}_{n,MMF}{J}_{0}\left({k}_{z,n}r\right)\,\,{e}^{i{{\beta }_{n}}{{L}_{MMF}}}\,,\] (1)

where L_MMF denotes the MMF length, N is the number of excited modes, and C_n,MMF are the modal decomposition coefficients. The resulting free-space field beyond the output facet can then be approximated as:

{E}_{freespace}\left(r,z\right)=\sum \limits_{n=1}^{N}{C}_{n,MMF}{J}_{0}\left({k}_{z,n}r\right)\,\,{e}^{i\left(\sqrt{{k}^{2}-k_{z,n}^{2}}z+{\beta }_{n}{L}_{MMF}\right)}\, ,\[{E}_{freespace}\left(r,z\right)=\sum \limits_{n=1}^{N}{C}_{n,MMF}{J}_{0}\left({k}_{z,n}r\right)\,\,{e}^{i\left(\sqrt{{k}^{2}-k_{z,n}^{2}}z+{\beta }_{n}{L}_{MMF}\right)}\, ,\] (2)

where z is the axial propagation distance from the MMF facet. Through the constructive interference of multiple Bessel-like fields, the output beam achieves near-diffraction-free propagation and sustains a well-defined transverse intensity profile over an extended axial range. As a greater number of guided modes participate in the superposition, the output field converges progressively toward a quasi-Bessel beam. Therefore, adopting a larger-core MMF, which accommodates more guided modes, is an effective approach to generate quasi-Bessel beams with an enhanced DOF.

4.2 Fabrication of the all-fiber OCT probe

As illustrated in Fig. 1(a), a section of UV-inscribed FBG with a center Bragg wavelength of 1550 nm was fusion spliced at the end of an SMF and precisely cleaved to a length of 8 mm. Subsequently, a segment of MMF (FG105LCA, Thorlabs, USA) was fusion spliced to the FBG end and cleaved to 2 mm, which enables an optimal modulation length for exciting the desired quasi-Bessel beam. Prior to encapsulation, the probe was sterilized by immersion in alcohol solution. It was then sealed within a 25-gauge hypodermic needle with inner and outer diameters of 300 μm and 500 μm, as shown in Fig. 1(d).

4.3 OCT system

This study employed a 1310 nm swept-source OCT system with an axial resolution of 14 μm. As illustrated in Fig. 2(a), the light emitted from a swept source (ranging from 1260 to 1360 nm and centered at 1310 nm) is divided by a 90/10 coupler (TW1300R2A1, Thorlabs, USA) and directed into the fiber probe in the sample arm and the mirror in the reference arm, respectively. Subsequently, the backscattered light collected by the probe is routed via a circulator to recombine with the reflected reference light. After passing through the 50/50 coupler (TW1300R5A1, Thorlabs, USA) and the 99/1 coupler, both signals are then fed into a balanced photodetector through port 2 and port 3 for interferometric demodulation. Finally, a custom-developed LabVIEW program is utilized for spectral demodulation to reconstruct the OCT images. To enable real-time temperature monitoring, an additional sensing arm for FBG demodulation was integrated into the system. Light emitted from an ASE source (HY-ASE-C-N, HOYATEK, China) enters the probe through a wavelength division multiplexer (WDM). The back-reflected light propagates through the sample arm and is extracted by a 99/1 fiber coupler, entering the OSA through port 1 for spectral analysis. Furthermore, the customized OCT software supports M-mode (time-mode) imaging, enabling continuous acquisition of A-scans over time with a refresh rate of 80 A-scans/s.

4.4 Preparation of the mice breast tumor model

To establish the orthotopic breast tumor model, 4T1 murine breast cancer cells were subcutaneously inoculated into the mammary fat pad of female BALB/c mice. When the tumor volume reached approximately 100 mm³, the mice were randomly assigned to experimental groups. The photothermal agent, tBuTTBD NPs-iRGD, AIEgen-based nanoparticles surface-functionalized with iRGD for tumor targeting, was administered via a single intravenous tail-vein injection²⁵. Time-dependent NIR-II fluorescence and photoacoustic imaging confirmed preferential nanoparticle accumulation at the tumor site within 12 h post-injection. Photothermal therapy was accordingly performed at this timepoint. As a comparison of temperature measurement, the thermal camera was positioned approximately 0.40 m from the mouse tumor and operated with an emissivity setting of 0.95, providing a spatial resolution of approximately 0.54 mm per pixel.

Competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Article History

Received Date: 2026-05-15

Accepted Date: 2026-09-04

Available Online: 2026-09-28

Corresponding Authors

• DJ Liu, E-mail: dejun.liu@szu.edu.cn

• CR Liao, E-mail: cliao@szu.edu.cn

• Issue Info: Vol. 1, No. 1, 260003-1–260003-12

• DOI: 10.67704/technol.2026.260003

Citation

Huang ZY, Liu DJ, Su CY et al. Dual-modal fiber needle probe for optical coherence tomography imaging and in situ temperature sensing. Technology 1, 260003 (2026)

Author Information

Ziyi Huang

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Corresponding author Dejun Liu, E-mail: dejun.liu@szu.edu.cn

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Chenyang Su

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Wenguang Zhang

School of Medicine, Wuhan University of Science and Technology, Wuhan 430065, China

Haoqiang Huang

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Yuxin Wang

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Lei Xu

School of Electronic and Communication Engineering, Shenzhen Polytechnic University, Shenzhen 518055, China

Dezhi Zhu

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Ying Wang

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Lei Song

Department of Cardiology, Fuwai Hospital, National Centre for Cardiovascular Diseases, National Clinical Research Centre for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100037, China

Junle Qu

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Liwei Liu

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Yiping Wang

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Corresponding author Changrui Liao, E-mail: cliao@szu.edu.cn

Shenzhen Key Laboratory of Ultrafast Laser Micro/Nano Manufacturing, Guangdong and Hong Kong Joint Research Centre for Optical Fiber Sensors, State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China

Shenzhen Key Laboratory of Photonic Devices and Sensing Systems for Internet of Things, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education/Guangdong Province, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China

Copyright & License

© The Author(s) 2026. Published by Technology Publishing.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

References

1. 1. Ran Y, Xu Z Y, Chen M F, et al. Fiber-optic theranostics (FOT): interstitial fiber-optic needles for cancer sensing and therapy. Adv Sci 9, e2200456 (2022). DOI: 10.1002/advs.202200456

2. 2. Deng K C, Tang Y, Xiao Y, et al. A biodegradable, flexible photonic patch for in vivo phototherapy. Nat Commun 14, 3069 (2023). DOI: 10.1038/s41467-023-38554-x

3. 3. Mansur A, Garg T, Shrigiriwar A, et al. Image-guided percutaneous ablation for primary and metastatic tumors. Diagnostics (Basel) 12, 1300 (2022). DOI: 10.3390/diagnostics12061300

4. 4. McDermott S, Gervais D A. Radiofrequency ablation of liver tumors. Semin Intervent Radiol 30, 49–55 (2013). DOI: 10.1055/s-0033-1333653

5. 5. Hernáandez J I, Cepeda M F J, Valdés F, et al. Microwave ablation: state-of-the-art review. Onco Targets Ther 8, 1627–1632 (2015). DOI: 10.2147/OTT.S81734

6. 6. Bodard S, Geevarghese R, Razakamanantsoa L, et al. Percutaneous cryoablation in soft tissue tumor management: an educational review. Insights Imaging 15, 278 (2024). DOI: 10.1186/s13244-024-01822-5

7. 7. Liu Y J, Bhattarai P, Dai Z F, et al. Photothermal therapy and photoacoustic imaging via nanotheranostics in fighting cancer. Chem Soc Rev 48, 2053–2108 (2019). DOI: 10.1039/C8CS00618K

8. 8. Overchuk M, Weersink R A, Wilson B C, et al. Photodynamic and photothermal therapies: synergy opportunities for nanomedicine. ACS Nano 17, 7979–8003 (2023). DOI: 10.1021/acsnano.3c00891

9. 9. Chen J H, Liu T S, Wang X L, et al. Display-based augmented reality navigation for CT-Guided thoracoabdominal punctures: phantom study of accuracy and workflow efficiency. Biomed Phys Eng Express 12, 025064 (2026). DOI: 10.1088/2057-1976/ae5827

10. 10. Kimbowa A, Pieters A, Tadayon P, et al. Advancements in needle visualization enhancement and localization methods in ultrasound: a literature review. Art Int Surg 4, 149–169 (2024). DOI: 10.20517/ais.2024.20

11. 11. Liang C-P, Wierwille J, Moreira T, et al. A forward-imaging needle-type OCT probe for image guided stereotactic procedures. Opt Express 19, 26283–26294 (2011). DOI: 10.1364/OE.19.026283

12. 12. Capon P K, Li J W, Horsfall A J, et al. A silk‐based functionalization architecture for single fiber imaging and sensing. Adv Funct Mater 32, 2010713 (2022). DOI: 10.1002/adfm.202010713

13. 13. Li J W, Warren-Smith S C, McLaughlin R A, et al. Single-fiber probes for combined sensing and imaging in biological tissue: recent developments and prospects. Biomed Opt Express 15, 2392–2405 (2024). DOI: 10.1364/BOE.517920

14. 14. Chen M H, Wang J P, Tan W J, et al. Miniaturized all fiber probe for optical coherence tomography and pH detection of biological tissue. J Biophotonics 14, e202000239 (2021). DOI: 10.1002/jbio.202000239

15. 15. Li J W, Schartner E, Musolino S, et al. Miniaturized single-fiber-based needle probe for combined imaging and sensing in deep tissue. Opt Lett 43, 1682–1685 (2018). DOI: 10.1364/OL.43.001682

16. 16. Krylov I V, Akasov R A, Rocheva V V, et al. Local overheating of biotissue labeled with upconversion nanoparticles under Yb(³⁺) resonance excitation. Front Chem 8, 00295 (2020). DOI: 10.3389/fchem.2020.00295

17. 17. Bianchi L, Mooney R, Cornejo Y, et al. Fiber Bragg grating sensors-based thermometry of gold nanorod-enhanced photothermal therapy in tumor model. IEEE Sens J 22, 11297–11306 (2022). DOI: 10.1109/JSEN.2021.3082042

18. 18. Yang Q C, Hu D M, Li Z S, , et al. Assembly tapered fiber Bragg grating tip with gold nanostars for heat generation and gradient temperature sensing. Opt Laser Technol 175, 110759 (2024). DOI: 10.1016/j.optlastec.2024.110759

19. 19. Lorenser D, Yang X, Kirk R W, et al. Ultrathin side-viewing needle probe for optical coherence tomography. Opt Lett 36, 3894–3896 (2011). DOI: 10.1364/OL.36.003894

20. 20. Zhu X, Schülzgen A, Li L, et al. Generation of controllable nondiffracting beams using multimode optical fibers. Appl Phys Lett 94, 201102 (2009). DOI: 10.1063/1.3138780

21. 21. Lorenser D, Yang X, Sampson D D. Accurate modeling and design of graded-index fiber probes for optical coherence tomography using the beam propagation method. IEEE Photonics J 5, 3900015 (2013). DOI: 10.1109/JPHOT.2013.2250939

22. 22. Su C Y, Li Z R, Liu D J, et al. Spatial beam intensity shaping based on an on-fiber 3D nanoprinted microlens. J Lightwave Technol 43, 3968–3973 (2025). DOI: 10.1109/JLT.2025.3529481

23. 23. Sun Z Q, Xiao X P, Zhao W L, et al. High-precision distributed temperature sensor based on phase-shifted FBGs interrogated by OFDR. J Lightwave Technol 43, 2357–2362 (2025). DOI: 10.1109/JLT.2024.3486925

24. 24. Yuan W, Mavadia-Shukla J, Xi J F, et al. Optimal operational conditions for supercontinuum-based ultrahigh-resolution endoscopic OCT imaging. Opt Lett 41, 250–253 (2016). DOI: 10.1364/OL.41.000250

25. 25. Zhang W G, Kang M M, Li X, et al. Interstitial optical fiber-mediated multimodal phototheranostics based on an aggregation-induced NIR-II emission luminogen for orthotopic breast cancer treatment. Adv Mater 36, e2406474 (2024). DOI: 10.1002/adma.202406474

26. 26. Ma M, Zhang Y, Gu N. Estimation the tumor temperature in magnetic nanoparticle hyperthermia by infrared thermography: phantom and numerical studies. J Therm Biol 76, 89–94 (2018). DOI: 10.1016/j.jtherbio.2018.07.004

27. 27. Ewald A J, Werb Z, Egeblad M. Monitoring of vital signs for long-term survival of mice under anesthesia. Cold Spring Harb Protoc 2011, pdb.prot5563 (2011). DOI: 10.1101/pdb.prot5563

28. 28. Grimaud J, Murthy V N. How to monitor breathing in laboratory rodents: a review of the current methods. J Neurophysiol 120, 624–632 (2018). DOI: 10.1152/jn.00708.2017

29. 29. Rivera D A, Buglione A E, Ray S E, et al. MousePZT: a simple, reliable, low-cost device for vital sign monitoring and respiratory gating in mice under anesthesia. PLoS One 19, e0299047 (2024). DOI: 10.1371/journal.pone.0299047

30. 30. Tan K M, Mazilu M, Chow T H, et al. In-fiber common-path optical coherence tomography using a conical-tip fiber. Opt Express 17, 2375–2384 (2009). DOI: 10.1364/OE.17.002375

31. 31. Kim J H, Han J H, Jeong J. Common-path optical coherence tomography using a conical-frustum-tip fiber probe. IEEE J Sel Top Quantum Electron 20, 6800407 (2014). DOI: 10.1109/JSTQE.2013.2277817

32. 32. Vairagi K, Gupta P, Tiwari U K, et al. Reflective axicon based energy-efficient extended depth of focus quasi-Bessel beam probe for common-path optical coherence tomography. Appl Opt 62, 511–517 (2023). DOI: 10.1364/AO.465544

33. 33. Gupta P, Vairagi K, Sharma V, et al. Tissue characterization using axicon probe-assisted common-path optical coherence tomography. Opt Express 32, 20194–20206 (2024). DOI: 10.1364/OE.508006

34. 34. Wang W, Wang G Y, Ma J, et al. Miniature all-fiber axicon probe with extended Bessel focus for optical coherence tomography. Opt Express 27, 358–366 (2019). DOI: 10.1364/OE.27.000358

35. 35. Wang W, Ma J, Guan B-O. Truncated conical-tip fiber probe for common-path optical coherence tomography with optimal sensitivity. J Lightwave Technol 40, 3974–3979 (2022). DOI: 10.1109/JLT.2022.3152617

36. 36. Moon S, Piao Z L, Kim C S, et al. Lens-free endoscopy probe for optical coherence tomography. Opt Lett 38, 2014–2016 (2013). DOI: 10.1364/OL.38.002014

37. 37. Lee J, Chae Y, Ahn Y C, et al. Ultra-thin and flexible endoscopy probe for optical coherence tomography based on stepwise transitional core fiber. Biomed Opt Express 6, 1782–1796 (2015). DOI: 10.1364/BOE.6.001782

38. 38. Ding Z H, Qiu J R, Shen Y, et al. Lens-free all-fiber probe with an optimized output beam for optical coherence tomography. Opt Lett 42, 2814–2817 (2017). DOI: 10.1364/OL.42.002814

39. 39. Qiu J R, Shen Y, Shangguan Z W, et al. All-fiber probe for optical coherence tomography with an extended depth of focus by a high-efficient fiber-based filter. Opt Commun 413, 276–282 (2018). DOI: 10.1016/j.optcom.2017.12.048

40. 40. Qiu J R, Han T, Liu Z Y, et al. Uniform focusing with an extended depth range and increased working distance for optical coherence tomography by an ultrathin monolith fiber probe. Opt Lett 45, 976–979 (2020). DOI: 10.1364/OL.383428

41. 41. Sapareto S A, Dewey W C. Thermal dose determination in cancer therapy. Int J Radiat Oncol Biol Phys 10, 787–800 (1984). DOI: 10.1016/0360-3016(84)90379-1

42. 42. Goldberg S N, Hahn P F, Tanabe K K, et al. Percutaneous radiofrequency tissue ablation: does perfusion-mediated tissue cooling limit coagulation necrosis?. J Vasc Interv Radiol 9, 101–111 (1998). DOI: 10.1016/S1051-0443(98)70491-9

43. 43. Lee J, Srinivasan V, Radhakrishnan H, et al. Motion correction for phase-resolved dynamic optical coherence tomography imaging of rodent cerebral cortex. Opt Express 19, 21258–21270 (2011). DOI: 10.1364/OE.19.021258

44. 44. Tavares C, Leitão C, Presti D L, et al. Respiratory and heart rate monitoring using an FBG 3D-printed wearable system. Biomed Opt Express 13, 2299–2311 (2022). DOI: 10.1364/BOE.452115

45. 45. Tosi D, Macchi E G, Gallati M, et al. Fiber-optic chirped FBG for distributed thermal monitoring of ex-vivo radiofrequency ablation of liver. Biomed Opt Express 5, 1799–1811 (2014). DOI: 10.1364/BOE.5.001799

46. 46. Unger M, Berger J, Melzer A. Robot-assisted image-guided interventions. Front Robot AI 8, 664622 (2021). DOI: 10.3389/frobt.2021.664622

47. 47. Yin B W, Piao Z L, Nishimiya K, et al. 3D cellular-resolution imaging in arteries using few-mode interferometry. Light Sci Appl 8, 104 (2019). DOI: 10.1038/s41377-019-0211-5

48. 48. Sheil C J, Wartak A, Spicer G L C, et al. Extended depth of focus by self-imaging wavefront division with the mirror tunnel. J Opt Soc Am A Opt Image Sci Vis 39, 711–725 (2022). DOI: 10.1364/JOSAA.448848

2026-0003_online.pdf