1 Introduction
Long-period fiber gratings (LPFGs) have long played an important role in such fields as high-temperature sensing and distributed sensing[1-2]. Compared with fiber Bragg gratings, the resonance peaks of LPFGs are more sensitive to different external environments. LPFGs have the advantages of mature fabrication techniques, immunity to electromagnetic interference, and high sensitivity. Therefore, they have been widely applied in strain[3-4], temperature[5-6], bending[7-8], refractive index (RI)[9-11], torsion[12-13], and virus[14] sensing. There are many methods for fabricating LPFGs. The most commonly used of these methods are femtosecond laser inscription[2,15], ultraviolet exposure[16] , and CO2 laser irradiation[17-18]. Femtosecond laser inscription has a robust anti-aging capability and large resonance wavelength depth.
In recent decades, the LPFGs with microchannel[19] and micropore[20] structures obtained by femtosecond laser inscription have been developed rapidly and widely applied in temperature, RI, and curvature sensing. Guo et al. proposed an LPFG based on periodic microchannels. First, a femtosecond laser is used to inscribe the periodic linear structures in single-mode fiber (SMF). Then, hydrofluoric acid is used to etch the laser-modified regions to obtain microchannels. This LPFG achieved highly sensitive measurements of environmental RI changes[19]. Lan et al. proposed a novel LPFG with internal micropores, which is highly robust. The femtosecond laser was first used to inscribe the periodic linear structures in the core of the SMF, and then micropores were formed by discharging in the region of the short lines[20]. Gong et al. developed a fiber Mach-Zehnder interferometer (MZI) sensor. The MZI was formed by fabricating a hollow ellipsoid in the fiber. The sensor can be used for bending sensing in an extensive curvature range[21]. From the works above, it can be concluded that the LPFG with microchannel and micropore structures exhibits good sensing characteristics and study values.
In this work, a novel LPFG based on periodic microchannels is proposed. The femtosecond laser first inscribes a series of line structures in the SMF cladding. Then, the periodic microchannels are formed by hydrofluoric acid etching of the femtosecond laser direct writing regions. The microchannels are filled with polydimethylsiloxane (PDMS) to obtain a better transmission spectrum. High sensitivity measurements of temperature, strain, bending, and refractive index have been achieved through this structure. This LPFG is low cost, robust, and has potential application value in multi-parameter sensing.
2 Experiment
2.1 Fabrication of LPFG
The experimental equipment of the LPFG fabrication system is shown in Fig. 1 (color online). The laser used in the system is the femtosecond laser of the Pharos series (PH1-10 W) manufactured by Light Conversion. The 1030 nm femtosecond laser is frequency doubled to 515 nm by using β-BaB2O4 (BBO) crystal. The repetition frequency is 10 kHz, and the pulse duration is 230 fs. After the femtolaser has passed through reflectors, a collimated beam, and a dichroic mirror, it is then focused into the fiber cladding by a 60× oil immersion objective (Olympus, NA = 1.42). The fiber is fixed on a three-axis translation platform controlled by a computer, which precisely controls the movement of the fiber. The laser direct writing process can be seen in real time on the computer screen through the charge-coupled device (CCD).

Figure 1.The LPFG fabrication system
In the experiment, the three-axis translation stage precisely controlled the movement of the fiber, and the computer precisely controlled the laser shutter. As shown in Fig. 2(a) (color online), the fiber moved perpendicular to the laser incident direction during fabricating to achieve a linear laser-modified structure. When the laser scanning position was close to the fiber’s core, the laser shutter was closed to position the laser modulation region in the fiber’s cladding. The laser scan length was approximately 56 μm. After scanning the first laser modulation region, the fiber was axially translated by 30 μm, and the above step was repeated, where the translation distance was the grating period. The LPFG has approximately 166 cycles and a total length of 10 mm.

Figure 2.(a) Schematic diagram of the periodic straight-line laser-modified structure by fs laser direct writing. (b) Microscope image of LPFG after HF etching. (c) Microscope image of LPFG filling with PDMS
Then, the microchannel structures were obtained by hydrofluoric acid etching of laser-modified regions. The periodic laser-modified linear fiber was immersed in a 20% hydrofluoric acid solution to achieve selective etching. During the above process, the etching was mainly carried out in the laser-modified region because the corrosion rate in the laser-modified region was more than two hundred times higher than that in the unmodified region[22]. To avoid high insertion loss, it was necessary to accurately control the etching time to ensure that the processed region was completely corroded without eroding into the fiber core. Through many comparative experiments, the best corrosion effect was found to be obtained by controlling the etching time at 20 minutes. At this duration, the spectrum was found to be optimal and there was no high insertion loss. Fig. 2(b) shows the microscopic image of the cladding etched LPFG. The approximate width of the microchannels measured at 8 μm.
After completing the above steps, the PDMS was filled into the corroded microchannels to improve the spectral quality. The PDMS solution was prepared with a 1∶10 ratio of S184-B and S184-A, and then the PDMS was dropped around the microchannels obtained by hydrofluoric acid etching. At the same time, the fiber was placed in the ultrasonic machine so that the PDMS entered the microchannels completely. The PDMS was solidified by a heating stage after completely entering the microchannels. Fig. 2(c) (color online) shows the microscopic image of the microchannel after filling the channels with PDMS.
2.2 Sensing principle
The LPFGs can achieve coupling between the core and cladding modes of the same direction transmission. According to the coupled mode theory, the resonant wavelength of the LPFG can be expressed as[23-25]:
$ {\lambda _{\mathrm{m}}} = \left( {n_{{\mathrm{eff}}}^{{\mathrm{co}}} - n_{{\mathrm{eff}}}^{{\mathrm{cl}}}} \right)\varLambda \quad,$ (1)
where ${\lambda _{\mathrm{m}}}$ is the resonant wavelength, $n_{{\mathrm{eff}}}^{{\mathrm{co}}}$ and $n_{{\mathrm{eff}}}^{{\mathrm{cl}}}$ are the effective refractive indicies of the core and cladding modes, respectively, and Λ is the grating period.
The resonant wavelength of the LPFG at different temperatures is expressed as[6]:
$ \frac{{{\mathrm{d}}{\lambda _{\mathrm{m}}}}}{{{\mathrm{d}}T}} = \left( {\frac{{{\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{co}}}}}{{{\mathrm{d}}T}} - \frac{{{\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{cl}}}}}{{{\mathrm{d}}T}}} \right)\varLambda + \left( {n_{{\mathrm{eff}}}^{{\mathrm{co}}} - n_{{\mathrm{eff}}}^{{\mathrm{cl}}}} \right)\frac{{{\mathrm{d}}\varLambda }}{{{\mathrm{d}}T}}.$ (2)
When the ambient temperature changes, ${\mathrm{d}}\varLambda /{\mathrm{d}}T$ represents the change in grating period caused by the fiber’s thermal expansion effect, ${\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{co}}}/{\mathrm{d}}T$ and ${\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{cl}}}/{\mathrm{d}}T$ represent the changes in the effective refractive index of the core and cladding modes caused by the thermal optical effect.
The resonant wavelength of the LPFG under the axial strain $\varepsilon $ is expressed as[4]:
$ \frac{{{\mathrm{d}}{\lambda _{\mathrm{m}}}}}{{{\mathrm{d}}\varepsilon }} = \left( {\frac{{{\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{co}}}}}{{{\mathrm{d}}\varepsilon }} - \frac{{{\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{cl}}}}}{{{\mathrm{d}}\varepsilon }}} \right)\varLambda + \left( {n_{{\mathrm{eff}}}^{{\mathrm{co}}} - n_{{\mathrm{eff}}}^{{\mathrm{cl}}}} \right)\frac{{{\mathrm{d}}\varLambda }}{{{\mathrm{d}}\varepsilon }} .$ (3)
When the axial strain is applied to the fiber, ${\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{co}}}/{\mathrm{d}}\varepsilon $ and ${\mathrm{d}}n_{{\mathrm{eff}}}^{{\mathrm{cl}}}/{\mathrm{d}}\varepsilon $ represent the changes in the effective refractive index of the core and cladding modes caused by the elastic optical effect.
As can be seen from formulas (2) and (3), since the refractive index of PDMS decreases when the temperature and strain increase, the resonant peak moves in the short wavelength direction in the temperature and strain sensing tests.
3 Results and discussion
After the device preparation was completed, the sensing characteristics of the device were investigated. Fig. 3 shows the initial transmission spectrum of the LPFG measured through an optical spectrum analyzer (OSA). The OSA is the model AQ6370D produced by Yokogawa. The highest resolution is 0.02 nm.

Figure 3.Initial transmission spectrum of LPFG
It can be observed from Fig. 3 that there are three distinct valleys in the initial transmission spectrum in the range of 1000 nm to 1700 nm, which are named Dip1, Dip2, and Dip3. The wavelengths of the Dips are 1225.65 nm, 1404.31 nm, and 1603.31 nm, respectively. The resonant wavelength Dip1 was observed in the following tests. Subsequently, a fiber sensing test system was established for temperature, strain, refractive index (RI), and bending sensing measurements.
3.1 Sensing test system
Fig. 4 shows the experimental equipment of the sensing test system. During the tests, light originated from the light source and entered the sensing device through the SMF. The SuperK COMPACT supercontinuum light source produced by NKT Photonics was used. It has an output wavelength of 400−2400 nm and an output power of 110 mW and is a non-polarized light source. The spectral changes during the process of sensing tests were recorded by the OSA.

Figure 4.Schematic diagram of the experimental setup for measuring (a) temperature, (b) strain and (c) bending
3.2 Temperature sensing
The spectral response to temperature is first carried out. Fig. 4(a) shows the setup of the temperature test. During the test, it was essential to keep the LPFG straight to avoid spectral changes due to bending. The LPFG was fixed in a tube furnace with a temperature resolution of 0.1 °C. The tube furnace was gradually heated from 26 °C to 90 °C in steps of 10 °C. At each temperature test point, the tube furnace temperature was maintained for sixty minutes to ensure that the environmental temperature of the LPFG had stabilized. The transmission spectra at different temperatures are recorded in Fig. 5(a) (color online). It shows that Dip1 blue shifts as the temperature increases. Dip1’s wavelength values at different temperatures are displayed in Fig. 5(b). The temperature sensitivity is −55.19 pm/°C in the range of 26 °C to 90 °C and the linear fit is 0.9939.

Figure 5.(a) Evolution of the transmission spectrum at different temperatures; (b) relationship between the resonance wavelength shift and the temperature
3.3 Strain sensing
During the strain sensing test, it was necessary to keep the environmental temperature of the LPFG stable to prevent temperature changes from affecting the strain test results. Fig. 4(b) shows the setup of the strain test. The axial tension was applied to the LPFG using a stress meter for the strain sensing test. The axial strain $\varepsilon $ was calculated by the equation:
$ \varepsilon = \frac{F}{{{\text{π}} {r^2}E}}\quad, $ (4)
where $F$ is the axial tension, $r$ is the cladding radius, and $E$ is the Young modulus of the silica fiber. The axial tension was gradually increased from 0 N to 1.4 N in steps of 0.2 N. The transmission spectra at different strains are recorded in Fig. 6(a) (color online). It can be observed that Dip1 blue shifts as the strain increases. Dip1’s wavelength values at different strains are displayed in Fig. 6(b). The strain sensitivity is −3.19 pm/με in the range of 0 to 1568 με and the linear fit is 0.9983.

Figure 6.(a) Evolution of the transmission spectrum at different strains; (b) relationship between resonance wavelength shift and strain
3.4 RI sensing
During the RI sensing test, the LPFG was placed in a hard plastic slot and kept straight. After fixing the LPFG, the solutions were injected into the plastic slot until they completely entered the microchannels. There are a total of 14 solutions with different RIs, which were made by mixing glycerin and water in different volume ratios. At each RI test point, the spectrum of the LPFG was recorded after it had stabilized. After recording the spectrum, the LPFG was repeatedly cleaned with deionized water and ethanol and finally dried with compressed air. When the LPFG was recovered, the above process was repeated to measure the other solutions with different RIs. The transmission spectra at different curvatures are recorded in Fig. 7(a) (color online). Dip1 red shifts as the RI increases. Dip1’s wavelength values at different RIs are displayed in Fig. 7(b) (color online). The RI sensitivities are 154.13 nm/RIU in the range of 1.333 to 1.399 and 540.81 nm/RIU in the range of 1.399 to 1.4275.

Figure 7.(a) Evolution of the transmission spectrum at different RI solutions; (b) relationship between resonance wavelength shift and RI
3.5 Bending sensing
In the bending sensing test, the LPFG was fixed on two fixtures, as shown in Fig. 4(c). The two fixtures were mounted on an adjustable stage with an accuracy of 0.1 mm. The LPFG was bent by moving the linear motion platform. The curvature C was calculated by the following equation[21]:
$ C = 2d/\left( {{d^2} + {L^2}} \right)\quad, $ (5)
where $L$ is the initial distance of the two linear motion platforms and $d$ is the movement distance of the two linear motion platforms.
During the sensing test, the curvature of the LPFG gradually increased from 0 to 3 m−1 in steps of 0.5 m−1. The transmission spectra at different curvatures are recorded in Fig. 8(a) (color online). The intensity of Dip1 increases as the curvature increases. Dip1’s wavelength values at different curvatures are displayed in Fig. 8(b). The bending sensitivity is 2.65 dB/m−1 and the linear fit is 0.9960.

Figure 8.(a) Evolution of the transmission spectrum at different curvatures; (b) relationship between resonance intensity shift and curvature
Table 1 compares the measurement parameters of sensors with different structures based on LPFG, and shows that the sensor developed in this paper exhibits relatively high sensitivity to temperature and can measure multiple parameters separately by controlling variables.

Table 1. Comparison of measurement parameters for different types of LPFG
Table 1. Comparison of measurement parameters for different types of LPFG
Year | Sensing structure | Temperature sensitivity | Strain sensitivity | RI sensitivity | Bending sensitivity | Reference | 2013 | Periodic microchannels | 9.95 pm/°C | −2.4 pm/με | −391 nm/RIU | − | [19] | 2017 | Hollow ellipsoid | − | − | − | 0.42 dB/m−1 | [21] | 2022 | Inner microholes | 13.06 pm/°C | −1.57 pm/με | − | − | [20] | 2022 | Micro air-channel | 12.1 pm/°C | − | 587.08 nm/RIU | − | [26] | 2023 | Taped two-mode fiber and PDMS | −0.412 nm/°C | −12.16 nm/MPa | − | − | [6] | 2023 | D-shape | 45 pm/°C | − | − | 17.6 nm/ m−1 | [27] | 2024 | Periodic microchannels on the cladding and PDMS | −55.19 pm/°C | −3.19 pm/με | 540.28 nm/RIU | 2.65 dB/m−1 | This work |
|
4 Conclusion
This paper develops a novel type of LPFG based on cladding periodic microchannels. The microchannels were obtained by femtosecond laser direct writing and hydrofluoric acid etching. The PDMS was filled into the periodic microchannels of the LPFG to obtain a better transmission spectrum. The LPFG sensor exhibits good sensing sensitivity. It has a temperature sensitivity of −55.19 pm/°C, a strain sensitivity of −3.19 pm/με, a maximum RI sensitivity of 540.28 nm/RIU, and a bending sensitivity of 2.65 dB/m−1. As a result of the compactness and high sensitivity of the device, this type of LPFG based on cladding refractive index modulation has potential applications in multi-parameter sensing and precision measurement.