The measurement of relative humidity (RH) is widely used including chemical applications, environmental monitoring, medical applications, industrial monitoring, agriculture, pharmaceuticals. Optical-fiber humidity sensors are attractive because of their good performance features, such as low electromagnetic interference, high sensitivity, adaptability to a harsh environment, and more [1_–3]. In recent years, many optical-fiber humidity sensors with different structures have been reported, including the fiber Bragg grating (FBG) [4], long-period grating (LPG) [5], Mach-Zehnder interferometer (MZI) [6, 7], Fabry-Perot interferometer (F-P) [8]. The optical-fiber Fabry-Perot sensor is widely used in many applications [9_–11] because of its simple structure, small size, good multiplexing ability, and high consistency [12, 13]. In recent years, optical-fiber F-P humidity sensors have been given more attention [14_–16].

In 2022, Sui [17] coated a layer of polyimide film upon the end of a single-mode-fiber—hollow-fiber structure to make the structure and polyimide film form a F-P cavity. The sensitivity of that system was −0.157 nm/%RH. Zhang et al. [18] proposed a fiber-optic humidity sensor based on a F-P interferometer with a hyaluronic acid (HA) and polyvinyl alcohol (PVA) composite membrane. The humidity characteristics of HA/PVA composite sensors with different mixing ratios were studied, with −233 pm/%RH humidity sensitivity. In 2021, Chen et al. [19] proposed a 3D-printed castle Fabry-Perot microcavity on the fiber tip for humidity sensing. By filling the FP microcavity with PVA, a sensitivity of 248.9 pm/%RH was obtained. In the same year, Wang et al. [20] used graphene quantum dots to fill an optical-fiber Fabry-Perot cavity formed by two single-mode optical fibers, and got 0.5655 nm/%RH sensitivity from 11% RH to 85% RH. Zhong et al. [21] constructed a Fabry-Perot interference structure by fusing hollow-core fiber to the end of a single-mode fiber and coating gelatin humidity-sensitive film to the end of the hollow-core fiber. The results showed that the sensitivity was 192 pm/%RH in the range of 20%–80% RH at 20 ℃. In 2020, Shrivastav et al. [22] proposed a humidity-sensing method based on a special microstructured fiber featuring a metal-organic framework (MOF). A humidity-sensor head was fabricated by depositing a chitosan polymer matrix in the pore of the MOF. The sensor had a maximum sensitivity of 81.05 pm/%RH in the range of 90 to 95% RH; Its linear-humidity-response range was from 70% to 95% RH. In 2019, Vaz et al. [23] made a fiber-optic F-P humidity sensor by stitching a small piece of hollow quartz tube into a single-mode fiber and applying polyvinylidene difluoride (PVDF) on the fiber’s end face in dip-coating fashion. The 32.54 pm/%RH relative-humidity sensitivity was obtained in the range of 50%_–70% RH.

Many reported F-P optical-fiber humidity sensors present low sensitivity, small linear-response range, weak response consistency, complex manufacturing process, or high-cost microstructured fiber. It is important to design an easily fabricated and low-cost optical-fiber humidity sensor with good performance. Many optical-fiber humidity sensors have been fabricated with humidity-sensitive materials by coating the periphery of the fiber, as a humidity-sensitive transducer. However, we use Nafion-PVA sol-gel and two normal single-mode optical fibers. By filling Nafion-PVA sol-gel into the optical-fiber F-P cavity, we develop an optical-fiber humidity sensor. The Nafion-PVA sol-gel acts not only as humidity-response transducer but also as a direct transmission medium, which enables a direct response to the humidity. Furthermore, Nafion-PVA sol-gel is transparent and adhesive, which is suitable for fabricating an optical-fiber F-P cavity. The sensor uses inexpensive materials and a simple fabrication process. Its high humidity sensitivity and good consistency are demonstrated in this paper.

2.1. Sensor Fabrication and Principle

2.1.1. Sensor Fabrication

Nafion is a polymer material developed by modifying Teflon. Its structure consists of a hydrophobic backbone similar to Teflon’s, and a side chain with a hydrophilic sulfonic acid group. The sulfonic acid group has strong hydrophilicity and swells dramatically when absorbing water, which gives Nafion a strong response to humidity. However, in experiments we find that Nafion solution is difficult to form into a flat and evenly distributed film at room temperature. To solve this problem, we mix PVA solution with Nafion solution. PVA is a hydrophilic substance with good film-forming properties at room temperature. A mixed solution is prepared from Nafion solution (5% mass fraction; DuPont Co., DE, USA) and PVA (70 mg/ml) in a certain volume proportion. Then we obtain Nafion-PVA sol-gel, and easily fabricate the Nafion-PVA film at room temperature. By filling the Nafion-PVA into the F-P cavity between two single-mode optical fibers, a high-sensitivity optical-fiber humidity sensor is easily fabricated from a simple structure.

The structure of the F-P optical-fiber sensor is shown in Fig. 1. The F-P cavity is composed of two ordinary single-mode optical fibers with flat end faces. Nafion-PVA film is filled into the cavity. Figure 1(a) is a schematic diagram of the structure, while Fig. 1(b) is an actual photo of the sensor head. The two fibers are well cut to get good reflection from the end faces. They are fixed on a fiber-adjusting frame and connected to a optical-fiber sensing analyzer (SM125; Micron Optics, GA, USA). The optical fibers are carefully adjusted to keep the two fiber ends as parallel as possible, to get a good F-P interference spectrum, which can be monitored by the SM125 sensing analyzer (Micron Optics). Then we add 3_–5 ml of Nafion-PVA mixed solution into the optical fiber F-P cavity with a dropper. To steadily introduce the Nafion-PVA mixed solution, a rectangular plastic gasket is placed under the fiber, as shown in Fig. 1(b). The fiber may be affected by liquid buoyancy after dropping the mixed solution, which causes deterioration of the F-P interference spectrum. In this case, the cavity should be adjusted again by the optical-fiber adjustment rack until the interference spectrum is satisfactory. Then the sensor is left to dry for 12 hours. Finally, the optical-fiber humidity sensor is obtained.

Figure 1. F-P sensor structure. (a) Schematic diagram of sensor structure. (b) Actual photograph of the sensor.

2.1.2. Humidity-response Principle

In this optical-fiber F-P cavity structure, the interference is formed by two reflected light beams, as Fig. 1(a) shows. The interference intensity can be expressed as

(1)$$I=I_0[R_1+R_2+2\sqrt{R_1{R}_2}\cos (\frac{2\pi L}{\lambda }+{\phi }_0)]$$

In Eq. 1, I₀ is the intensity of the incident light, R₁ and R₂ are the reflectance values for the two fiber ends, λ is the wavelength of the light, and φ₀ is the phase constant. L represents the optical-path difference of the two reflected beams, and

(2)$$L=2nd$$

Here n is the effective refractive index of the Nafion-PVA material, and d represents the F-P cavity’s length.

When the phase satisfies the interference phase condition, the maximum value of light intensity is obtained. If the peak wavelength of the reflection interference spectrum is λ_m, the expression for this peak wavelength is

(3)$${\lambda }_m=\frac{2nd}{m}$$

where m is an integer. Obviously λ_m is determined by the length d and the effective refractive index n of the F-P cavity. Thus the drift of the interference spectrum is

(4)$$\frac{\Delta {\lambda }_m}{{\lambda }_m}=\frac{\Delta n}{n}+\frac{\Delta d}{d}$$

When the humidity varies, the refractive index of the Nafion-PVA film changes. Also, Nafion-PVA film swells with increasing humidity, which causes the cavity length to change. Therefore the interference spectrum shifts with external humidity change.

The free spectral range (FSR) is the wavelength difference between two adjacent peaks or troughs of the spectral curve, which can show the density of the spectral curve, an important parameter of F-P interference spectrum. The definition of FSR is

(5)$$\text{FSR}={\lambda }_m-{\lambda }_{m+1}$$

where λ_m and λ_m+1 are the adjacent peak wavelengths corresponding to the reflectance spectrum. According to Eqs. 3 and 5, the FSR can be expressed as

(6)$$\text{FSR}=\frac{{\lambda }_{m+1}{\lambda }_m}{2nd}$$

The FSR also changes with the humidity, because of the effective refractive index n and cavity length d, so optical-path difference (OPD) and FSR are both related to the external humidity.

2.2. Experimental System

For humidity-sensing experiments, the setup is shown in Fig. 2. The experimental system includes a SM125 optical-fiber sensing analyzer (Micron Optics), computer data acquisition and demodulation module, humidity sensor, and humidity bottle with holder for probe support. The SM125, which can be used as a light source and spectrum-measurement device, has an output power of 18 mW, wavelength range of 1,510_–1,590 nm, scanning frequency of 2 Hz, wavelength accuracy and stability of 1 pm.

Figure 2. Experimental system for measuring relative humidity.

The light emitted by the SM125 is transmitted to the optical-fiber F-P sensor, and the F-P interference spectrum is obtained. Then the interference data are transmitted to the computer. When the ambient humidity changes, the relationship between spectrum shift and ambient humidity can be measured.

In the experiments, humidity bottles with different saturated salt solutions provide an environment with RH of 29%–72%. Before each experiment, we use a hygrometer to calibrate the humidity value within the bottle. The optical-fiber sensor is put into different humidity bottles, and the humidity-response spectra are monitored by the SM125. To ensure that the sensing head does not touch the saturated salt solution, the sensor head is fixed on a holder of suitable height for support inside the humidity bottle.

3.1. Humidity-response Measurement

In the experiments, the optical-fiber F-P sensor is put into different humidity bottles to observe the peak-wavelength drift of the interference spectrum. We mixed Nafion (5% solution; DuPont Co.) and 70 mg/ml PVA solution in a 6:1 volume ratio. The initial ambient humidity measured is 29%, at ambient temperature of 18 ℃. The response spectrum is monitored for 600 s at given humidity value, as shown in Fig. 3; Good stability of the spectral response is observed. According to the experimental data, the maximum peak-wavelength drift is ±0.3 nm. Spanning the range of 29%–72% RH, the sensor’s response is investigated at 29%, 43%, 54%, 66%, and 72% RH. With increasing humidity, the peak wavelength drifts to the right because the Nafion-PVA film interacts with water vapor. As a result, the Nafion-PVA’s refractive index and the cavity’s length both change. Thus the optical-path difference between the two reflected beams in the F-P cavity varies, which finally causes motion of the interference spectrum. The interference spectra are presented in Fig. 4.

Figure 3. Stability of the humidity-response spectrum.

Figure 4. Humidity-response spectra.

The relationship between peak-wavelength shift and relative-humidity change can be obtained from the interference spectra, and is displayed in Fig. 5. There is a good linear relationship between peak-wavelength shift and RH. The linear-fitting results show that the sensitivity of the sensor is 2.25 nm/%RH, with a linear correlation coefficient of 0.9997. The sensitivity is high, and some peak-wavelength drift is so large as to be beyond the detection range of the instrument. To avoid affecting the results, extra attention should be paid to the peak drift during the experiment, and the spectral information should be recorded in time. Here multiple peaks are monitored to obtain the relationship between humidity and peak-wavelength drift.

Figure 5. Wavelength shift for different humidity values, measured for both increasing and decreasing humidity.

To evaluate the consistency of this humidity response, the experimental wavelength-shift curves were obtained with both increasing and decreasing humidity, as Fig. 5 shows. The response curves for rising and falling humidity nearly coincide, with linear correlation coefficients of 0.99967 and 0.99595 respectively. Obviously the curves show good humidity-response consistency.

According to the spectral data, the corresponding curve for FSR and humidity is also obtained, as Fig. 6 shows. Due to the large wavelength drift and large range of humidity, multiple peaks are tracked to measure FSR, because of the limitations of the SM125’s measurement range. Even so, the high sensitivity and wide humidity range can render the FSR change impossible to measure. Therefore, here we measure a typical humidity sensor over the reduced range of 51%–75% RH. As Fig. 6 indicates, the linear-fitting results show the sensitivity of the sensor to be 0.03321 nm/%RH, with a linear correlation coefficient of 0.995.

Figure 6. FSR-response curve. FSR, free spectral range. 1:6 PVA:Nafion volume ratio, 51.9 μm initial cavity length.

We also found the relationship between humidity and optical-path difference, as shown in Fig. 7. Using OPD to evaluate the response, the sensitivity from this figurerises to 80.3 nm/%RH, with a linear correlation coefficient of 0.971. This method can effectively amplify the sensitivity.

Figure 7. Optical-path difference (OPD) responding curve.

3.2. Influences on Humidity Response

To further investigate the characteristics of the sensor’s humidity response, experiments were performed using sensors with different PVA:Nafion ratios, and using different cavity lengths. The response curves are presented in Fig. 8. Figure 8(a) shows the response curves for three sensors with different values of the PVA:Nafion volume ratio (1:4, 1:5, and 1:6 respectively). The cavity length for these three sensors is about 90 μm. Figure 8(b) shows the response curves for three sensors with differing initial length of the F-P cavity. These three sensors used the same PVA:Nafion volume ratio of 1:6.

Figure 8. Response curves for different material solutions and cavity lengths. (a) Different material solutions, (b) different initial cavity lengths.

According to Fig. 8(a), all three sensors show good linear humidity response, with 0.982, 0.988, and 0.999 correlation coefficients respectively. For the three values of material ratio, the sensitivity is 1.01 nm/%RH, 1.48 nm/%RH, and 2.25 nm/%RH respectively. With increasing Nafion ratio, the sensitivity of the humidity respondse increases. Figure 8(b) shows the response curves for three sensors with initial cavity lengths of 49.7 μm, 69.3 μm, and 93.2 μm respectively. The cavity length can be measured and controlled by a microscope, through careful adjustment and SM125 monitoring. These experiments also show good linear humidity response for all three sensors. However the sensitivity changes only slightly for different initial cavity lengths. The influence of the PVA:Nafion volume ratio is much more obvious.

3.3. Comparison to Other Optical-fiber Humidity Sensors

Table 1 shows a comparison between our work and some other optical-fiber humidity sensors. Compared to these others, our sensor has high response sensitivity and good response linearity, with easy fabrication and low cost. Compared to references [24_–26], although their sensors can work at higher humidity values, our sensor’s humidity response is more sensitive. The humidity-response range of reference [27] is smaller than in our work. The sensors of references [27_–30] have lower sensitivity, complex fabrication, or a high-cost microstructured fiber. The sensor of reference [31] has high sensitivity, but its fabrication is complex.

Table 1 . Comparison to other optical-fiber humidity sensors.

Reference	Type	Sensitive Materials	RH Range (%)	Sensitivity
24	Balloon-like Fiber	Gold Nanoparticles	35–95	−0.571 nm/%RH
25	Mach-Zehnder Interferometer	Reduced Graphene Oxide	45–95	0.2768 nm/%RH
26	Michelson Interferometer	Methylcellulose	55–100	0.133 nm/%RH
27	Tapered Optical Fiber	Silk Fibroin-LiBr Composite	29–39	−2.3 nm/%RH
28	Optical-fiber Grating	Graphene Oxide	30–80	18.5 pm/%RH
29	F-P (UV Drawing-curing)	Poly (methyl methacrylate)	61–71	659.67 pm/%RH
30	F-P (Etched Fiber)	C60-THAM	30–65	0.42 nm/%RH
31	F-P (Glass-capillary Heating)	Nafion	30–85	3.78 nm/%RH
Our Work	F-P (Common Fiber)	Nafion-PVA	29–72	2.25 nm/%RH

An optical-fiber F-P humidity sensor filled with Nafion-PVA sol-gel was fabricated and experimentally investigated. The sensor exhibited a good linear response (0.9997 linear correlation coefficient), with a sensitivity of 2.25 nm/%RH sensitivity from 29% to 72% RH. Its good response consistency was also verified in experiments with increasing and decreasing humidity. Experiments with different PVA:Nafion volume ratios and cavity lengths also showed that the sensor had good linear-response characteristics and high sensitivity. The humidity-response sensitivity increased with higher Nafion content (lower PVA:Nafion volume ratio). However the sensitivity change caused by changing initial length of the F-P cavity was little. The influence of the PVA:Nafion volume ratio on sensitivity was more obvious than the influence of cavity initial length. This sensor offers good humidity-response performance and an easy fabrication process.

The authors declare no conflicts of interest.

Data underlying the results presented in this paper are not publicly available at the time of publication, but may be obtained from the authors upon reasonable request.

National Natural Science Foundation of China (No. 61890964); Key Deployment Project of Ocean Science Research Center of Chinese Academy of Sciences (COMS2020J11); National Training Program of Innovation and Entrepreneurship for Undergraduates (No. 202111025 and 202211033).