Open Access
Add to BibSonomy
Abstract
Although numerous efforts have been dedicated towards developing fiber sensors with high performances, challenges still remain in achieving high-quality temperature sensors with high sensitivity, large measurement range and high stability. This study proposes a compact fiber optic temperature sensor based on PDMS-coated Mach-Zehnder interferometer (MZI) combined with FBG, and it can realize both high-sensitivity and large-range temperature measurement. The MZI is based on Thin No-Core Fiber (TNCF) with lateral-offset. Owing to the high refractive index sensitivity of MZI and the high thermo-optic coefficient of PDMS, the sensor can achieve a high temperature sensitivity (>10 nm/°C). Besides, by optimizing the TNCF length, the cascaded FBG can be used to locate different temperature intervals in units of approximately 10 °C, and therefore the detectable temperature range is largely extended. The experimental test demonstrates that the average sensitivities of 11.19 nm/°C, 8.53 nm/°C, 7.76 nm/°C, 7.27 nm/°C are achieved at the temperature around 30 °C, 40 °C, 50 °C and 60 °C, and it shows excellent consistency and repeatability during the thermal cycle tests.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
High-accuracy temperature detection has been widely applied in many fields such as marine environment [1], ground-coupled heat pump [2], pipeline monitoring [3]. When compared with the electric sensors, the fiber optic sensors own several advantages such as high sensitivity, small size, low cost, immune to electromagnetic interference, easiness to form the distributed sensing network. The traditional fiber optic temperature sensor mainly relates to the fiber Bragg grating (FBG) [4] or LFPG [5], but their relatively low sensitivities limit the actual applications. In recent years, a great number of interferometric fiber optic temperature sensors with novel design have been reported, such as multimode-core fibers [6], few-mode fiber [7], photonic crystal fibers [8], side-polish fibers [9], taper fibers [10], exposed-core fibers [11] etc. These configurations exhibit excellent temperature sensitivity performances at the expense of complex fabrication technique and high cost.
Besides, the refractive index (RI) sensors can be also converted into temperature sensing by combining the high thermo-optical coefficient material [12]. As one of the typical RI sensors, the fiber optic Mach-Zehnder interferometer (MZI) with lateral-offset splicing has attracted great interest due to its high sensitivity, compact structure and low cost. Several types of configurations have been previously reported [13,14]. Then, by covering a layer of high thermo-optical coefficient material, such as Polydimethylsiloxane (PDMS), the temperature measurement can be realized. However, considering the MZI configurations, their detectable temperature range is strictly limited by the free spectral range (FSR) of interference spectrum. Generally speaking, the higher the sensitivity is, the smaller the temperature range is. In the previously reported work, the temperature measurement range is limited to several degrees when ultra-high sensitivity (>10 nm/°C) is obtained [15,16]. Hence, temperature sensing with both high sensitivity and large range is desired.
In this study, we propose a fiber optic temperature sensor based on PDMS-coated MZI combined with FBG, and it is able to realize both high-sensitivity and large-range temperature measurements. The fiber optic MZI is based on TNCF (diameter of 61.5 $\mathrm{\mu }\textrm{m}$) with lateral-offset splicing. Owing to high RI sensitivity of MZI and high thermo-optic coefficient of PDMS, a high temperature sensitivity can be obtained. Besides, by optimizing the TNCF length, the cascaded FBG can be used to locate different temperature intervals in units of approximately 10 °C, and therefore the detectable temperature range is largely extended. Firstly, the fabrication process and sensing principle of the sensor are theoretically described. After the sensor is experimentally fabricated, the FSRs, sensitivities, and Figure of Merits (FOM) around different temperatures are discussed. Finally, the sensor consistency and repeatability during thermal cycle tests are investigated. The proposed temperature sensor with both high sensitivity and large range has potential applications in various fields such as environment monitoring, food safety detection.
2. Sensor fabrication and principle
The structural diagram of the fiber optic temperature sensor is shown in Fig. 1 (a). It consists of PDMS coated MZI and cascaded FBG. Among them, the PDMS coated MZI is designed as a cascaded fiber structure, which mainly consists of lead-in SMF (diameter: 9/125 $\rm{\mu }\textrm{m}$), lead-in NCF (diameter: 125$\mathrm{\mu }\textrm{m}$), sensing TNCF (diameter: 61.5 $\mathrm{\mu }\textrm{m}$), lead-out NCF (diameter: 125 $\mathrm{\mu }\textrm{m}$), lead-out SMF (diameter: 9/125 $\mathrm{\mu }\textrm{m}$). The 61.5 $\mathrm{\mu }\textrm{m}$ TNCF is spliced between two 125 $\mathrm{\mu }\textrm{m}$ NCF with a lateral offset distance of 30 $\mathrm{\mu }\textrm{m}$, so that the transmitted light power can be evenly divided into the sensing/reference arms of the MZI, and the best interference fringe can be obtained. The fabrication procedure of the sensor is shown in Fig. 1(b) and described as below: (1) A segment of SMF is spliced with the NCF (diameter: 125$\mathrm{\mu }\textrm{m}$) by using a fiber fusion splicer (Fujikura FSM-100P+, Japan), and only 1 mm length of NCF is retained. The fusion splicing parameters of this step are set as: arc power∼308 units, arc duration∼800 ms. (2) The cascaded SMF-NCF is then spliced with the TNCF with intentional lateral-offset of 30$\mathrm{\mu }\textrm{m}$ in Y direction, while they are aligned along X direction. The fusion splicing parameters of this step are set as: arc power∼ 180 units, arc duration∼200 ms. (3) The TNCF is cleaved precisely under a microscope, and TNCF length of 340$\mathrm{\mu }\textrm{m}$ is retained. A detailed discussion of TNCF length will be presented later. (4) A copy of SMF-NCF is spliced with the other end of sensing NCF. In this step, the fusion splicing parameters are the same as the second step. (5) The PDMS layer is covered onto the NCF surface. The composite structure is heated at 80 °C and then cooled down until it is solidified completely. The detailed elaboration process of PDMS is already illustrated in the authors’ previous study [17]. (6) The PDMS-coated MZI is spliced with an FBG (central wavelength: 1570 nm. Reflectivity: 93%, 3 dB bandwidth: 1 nm) to form the final sensing structure.
Fig. 1. Diagram of fiber optic temperature sensor (SMF: single mode fiber, NCF: no-core fiber, TNCF: thin no-core fiber, PDMS: Polydimethylsiloxane, FBG: fiber Bragg grating).
After the sensor is fabricated successfully, it is placed onto a tiny purpose-built heating plate for temperature test. During the temperature test, two fiber ends of the sensor device are connected with Broad Bandwidth Source (BBS, Fiberlake Co.,Ltd, WBB400008SFA) and the Optical Spectrum Analyzer (OSA, Yokogawa, AQ6370D) for transmission spectrum acquisition, and a K-type thermocouple is used for temperature calibration.
In this study, the PDMS-coated MZI and the FBG are cascaded to achieve both high-sensitivity and large-range. Firstly, the sensing principle of PDMS-coated MZI is briefly described. When the light transmits through the interface between SMF and lead-in NCF, a large number of high-order modes are excited. The lead-in NCF acts as a ‘light beam expander’ to improve the power coupling efficiency. Then, the high-order modes continue to transmit and part of them are coupled into the sensing TNCF, which form the reference arm of MZI. Meanwhile, the rest part of high-order modes transmit along the exposed cavity which is filled by the PDMS, and they form the sensing arm of MZI. Two parts of light are recoupled back into the lead-out NCF and the interference phenomenon occurs. The RI difference between PDMS and TNCF is significantly larger than that of different high-order modes. Hence, the proposed interferometer can be approximately simplified as a two-beam interference. Meanwhile, a high RI sensitivity is obtained. Besides, owing to the excellent thermo-optic coefficient of PDMS (∼-4.2 × 10−4 RIU/°C [18]), the external temperature variation can be converted into the RI changes of PDMS with high efficiency, hence a high temperature sensitivity can realize after covering the PDMS layer onto the MZI. For the proposed PDMS-coated MZI, the resonant wavelength of the interference dip is expressed as:
where $\mathrm{\Delta }{n_{eff}}$ is RI difference between the sensing arm and the reference arm, L is the TNCF length, m is interference order.Besides, the temperature-dependent RI (nT) of the PDMS can be expressed as:
where n0 is the initial RI of PDMS at ambient temperature, dn/dT is the thermo-optic coefficient, T is the actual temperature. Combining Eq. (1) and Eq. (2), the theoretical temperature sensitivity of the sensor can be expresses as: Furthermore, the free spectral range (FSR) describes the wavelength spacing between the adjacent interference dips,which is expressed as: Although the PDMS-coated MZI can realize high-sensitivity temperature measurement, its temperature detection range is limited due to the FSR. For example, assuming that nNCF=1.465, npdms=1.412 (T=25 °C) [18], dn/dT=-4.2 × 10−4 RIU/°C, L=340$\mathrm{\mu }\textrm{m}$, $m=13$, then, the theoretical temperature sensitivity is estimated as 10.58 nm/°C according to Eq.(3). On the other hand, the FSR (m=12∼13) at 25 °C is approximately estimated as 106.8 nm according to Eq.(1), which signifies that the detectable temperature range is only 106.8/10.58 = 10.09 °C. When the temperature variation range exceeds this value, the resonant wavelength of the former interference dip will coincide with that of the latter interference dip, which makes the temperature sensing mechanism unavailable. Hence, the detectable temperature range is limited.To solve this problem, an FBG is cascaded with the PDMS-coated MZI to extend the temperature measurement range. To be more specific, the central wavelength of FBG is used to divide different temperature intervals due to its relatively low temperature sensitivity. Here, the temperature interval is defined as the ratio between FSR and the corresponding dip sensitivity. To simplify the analysis, in this study, the temperature interval is designed around 10 °C, e.g. 20∼30 °C, 30∼40 °C, etc. Hence, the temperature demodulation principle of the proposed sensor can be described as below (shown in Fig. 2). Firstly, we fix the specific temperature interval by locating the Bragg wavelength. Then, focusing on one of the interference dips, we can obtain the corresponding resonant wavelength shift when compared with the transmission spectrum of the initial temperature of the interval (e.g. spectrum at 30 °C for the interval 30∼40 °C). Finally, based on the temperature sensitivity of the selected interference dip, we can demodulate the temperature. Hence, the detectable temperature range is largely extended by using this method.
Fig. 2. Diagram of temperature sensing principle (taking temperature interval of 30∼40 °C as example).
3. Optimizations of sensor parameters
From Eq. (1), it is seen that the length of lead-in/out NCF has no direct effect on the interference wavelength location. However, different from the core/cladding structure of SMF, the evanescent wave transmitting through NCF contacts directly with the surrounding environment, which results in a greater optical power loss. Hence, the lengths of lead-in/out NCF usually do not exceed 1 mm. Besides, in this study, the sensing TNCF with diameter of 61.5$\mathrm{\mu }\textrm{m}$ is selected. Hence, the lateral offset distance in Y direction is set as 30 $\mathrm{\mu }\textrm{m}$ to ensure that the optical power can be evenly divided into both sensing and reference arms. Under this condition, the best interference fringe visibility can be achieved. Finally, Fig. 3 shows the theoretical temperature interval versus the sensing TNCF length according to Eq.(1)-(4). We can observe that the theoretical temperature interval decreases as the TNCF length increases. As mentioned previously, the theoretical temperature interval is designed around 10 °C to simplify the Bragg wavelength localization. Hence, the sensing TNCF length should be approximately around 347 $\mathrm{\mu }\textrm{m}$. In this study, as mentioned previously, the actual TNCF length of 340$\mathrm{\mu }\textrm{m}$ is retained for the fabricated sensor.
The previous paragraphs present the simplified discussions of temperature interval to explain the purpose of FBG. However, the situation is more complex when the testing temperature varies. For the actual temperature sensing applications, the spectrum range used for observation is usually fixed, e.g. 1300 nm∼1600 nm, and an interference dip with a specific order (i.e. ‘m’ in Eq.(1)) is selected to observe the resonant wavelength shifts. According to Eq. (1)-(2), the resonant wavelength shows the redshift as the temperature increases. Since the temperature sensitivity of PDMS-coated MZI is larger than 10 nm/°C, the resonant wavelength may shift out of the observed spectrum range when the temperature increases to a certain value. In this case, the next interference dip (i.e. ‘m+1’) will be focused for spectral observation. Hence, we can imagine that the actual temperature sensitivity of the ‘focused interference dip’ decreases gradually as the interference order increases (‘m’->’m+1’) according to Eq. (3). Besides, assuming that the observed wavelength remains the same, the increase of testing temperature also results in the decrease of FSR according to Eq. (2) and (4). From the above analyses, both FSR and temperature sensitivity decrease as the testing temperature increases. Hence, if the testing temperature variation is added into consideration, the theoretical temperature interval should be thoroughly investigated. Table 1 shows the theoretical results of FSRs, sensitivities and the corresponding temperature interval under different testing temperatures. To better compare with the following experimentally fabricated sensor device, the sensing NCF length is also set as 340 $\mathrm{\mu }\textrm{m}$ for theoretical calculations. We can observe that, although both FSR and sensitivity decrease as the testing temperature increases from 30 °C to 60 °C, the temperature interval always keeps around 10 °C. This result confirms the feasibility of sensor design.
Table 1. Theoretical FSR, sensitivity and corresponding temperature interval.
4. Experimental results and discussions
Firstly, Fig. 4 shows the initial transmission spectrum of the sensor at room temperature (T=26 °C). The three obvious interference dips generated from interferometer are observed. According to Eq. (1), they correspond to the interference orders of 14, 13 and 12. Besides, the approximate insertion loss of the sensor is 26.95 dBm, the fringe depths and Full Width at Half Maxima (FWHM) of the three dips are 10.25 dB/17.8 nm, 26.38 dB/3.5 nm and 11.99 dB/19.1 nm. At the same time, the Bragg wavelength is observed locating around 1570 nm.
Then, several typical transmission spectra (T=30, 40, 50, 60 °C) are selected as examples to analyze the wavelength shifts caused by the interference dips and FBG respectively, and the results are shown in Fig. 5. Firstly, the Bragg wavelength shift is discussed. We can observe that the Bragg wavelength shifts are insignificant when temperature increases from 30 °C to 60 °C, and the averaged sensitivity is estimated as 12 pm/°C. As mentioned previously, the FBG is used to locate different temperature intervals. That is to say, the temperature range of 30∼40 °C covers the wavelength varying from 1570.04 nm to 1570.16 nm, and the temperature range of 40∼50 °C covers the wavelength varying from 1570.16 nm to 1570.28 nm, and so on. Hence, the temperature interval can be determined by locating the Bragg wavelength.
Fig. 5. (a) Transmission spectra of the sensor at the temperatures of 30 °C, 40 °C, 50 °C, 60 °C. (b) Enlarged view of Bragg wavelengths.
In the following part, we still focus on Fig. 5, and the interference dips are discussed. The average FSRs of the four transmission spectra (T30∼T60) are 102.6 nm, 95.4 nm, 88.6 nm, 82.2 nm respectively, while their corresponding theoretical average values are determined to be 103.4 nm, 96.4 nm, 90.3 nm, 84.9 nm according to Eq.(4). The differences of FSRs between experiments and theoretical predictions are mainly caused by the sensing TNCF length. It is rather difficult to obtain a high accuracy value since the TNCF is measured under the microscope, and its total length is only several hundred microns.
Next, Fig. 6(a) shows the transmission spectra of the sensor at the temperatures of 30 °C, 40 °C, 50 °C and 60°C. We observe that all the resonant wavelengths of the interference dips vary periodically as temperature increases, and the corresponding temperature sensitivities are shown in Fig. 6(b). Based on Eq. (3), the theoretical dip sensitivity should decrease as the interference order increases. That is to say, dip C should present the highest temperature sensitivity while dip A should present the lowest. However, in Fig. 6(b), the experimental sensitivities of dip A, B and C do not vary monotonously. This phenomenon may be explained by the non-uniform coating of the PDMS layer during the solidification process, which is rather difficult to intentionally control. When the measurement light of the interferometer transmits through the PDMS layer, the non-uniform coating layer may affect the specific optical propagation paths of the different interference modes, which finally affects the interference sensitivity. Besides, the average temperature sensitivities of 3 dips are 11.19 nm/°C at 30 °C, 8.53 nm/°C at 40 °C, 7.76 nm/°C at 50 °C, 7.27 nm/°C at 60 °C. These values are slightly lower than the theoretical results shown in table1, but they still show good agreement. The difference between them may be explained by the selected theoretical values of the thermo-optic coefficient of PDMS and the TNCF length. In addition, the theoretical result in Table 1 corresponds to the sensitivity of a single interference dip (i.e. m=13, 14, 15 or 16), while the experimental result corresponds to the average sensitivity of three dips (i.e. dip A∼C), this also leads to the differences in sensitivity. Finally, although dip B shows a relative low sensitivity, its presents a high Figure of Merit (FOM, defined as the ratio between sensitivity and 3 dB bandwidth) due to its narrow dip bandwidth (<5 nm), and the corresponding FOM values are determined to be 1.92 ${{^{\circ}{\textrm C}}^{ - 1}}$ (30 °C), 1.83 ${{^{\circ}{\textrm C}}^{ - 1}}$ (40 °C), 1.75 ${{^{\circ}{\textrm C}}^{ - 1}}$ (50 °C), 1.37 ${{^{\circ}{\textrm {C}}}^{ - 1}}$ (60 °C) respectively.
Fig. 6. Transmission spectra and the corresponding sensitivities of the sensor around different temperatures (a) 30 °C. (b) 40 °C. (c) 50 °C. (d) 60 °C.
To realize high accuracy sensing, the consistency and repeatability of the sensor should be discussed. Figure 7 shows the thermal behavior of the sensor during two continuous heating-cooling cycles. In this study, the resonant wavelengths between the first heating cycle and the second heating cycle are compared to evaluate the sensor repeatability, as shown in Fig. 7(a). The corresponding comparison result during the cooling cycles is shown in Fig. 7(b). Furthermore, the resonant wavelengths between the first heating cycle and first cooling cycle are compared to evaluate the sensor consistency during heating-cooling process, as shown in Fig. 7(c). From the above analysis, we know that dip B shows the lowest temperature sensitivity of three dips. Correspondingly, dip B also represents the worst resolution or the largest measurement error. Hence, theoretically speaking, dip A and dip C should represent better consistency and repeatability than dip B during the thermal cycle process. Here, dip B is selected as an example for the following analysis, several temperature points varying from 30 °C to 50 °C with an increment of 5 °C are selected. Besides, the resonant wavelengths with different interference orders are observed to ensure that the detected wavelength is always around 1400 nm. In Fig. 7(a)-(c), we observe that the average wavelength deviations are 0.22 nm, 0.58 nm and 0.84 nm respectively. Since the averaged sensitivity of dip B is calculated as 6.76 nm/°C, the corresponding temperature measurement errors are estimated as 0.032 °C, 0.085 °C, 0.12 °C. Then, the maximum deviation rate is determined to be 0.12/(50-30) = 0.6%. Hence, the presented results confirm the excellent consistency and repeatability of the proposed temperature sensor.
Fig. 7. Sensor consistency and repeatability test results (a) dip wavelength comparison between the first and second heating cycle; (b) dip wavelength comparison between the first and second cooling cycle; (c) dip wavelength comparison between the first heating and cooling cycle; (‘m’ is the interference order).
Generally speaking, the performance of PDMS becomes worse when the temperature exceeds 200 °C, hence the theoretical maximum temperature range can achieve as high as 200 °C. However, the sensitivity of the sensor decreases significantly as the temperature increases, e.g. the sensitivity value at 60°C is almost half of that at 30 °C, and it becomes less competitive when compared with the other sensors. Hence, the temperature sensing higher than 60 °C is not discussed in this study.
Finally, a comprehensive performance comparison between different fiber optic temperature sensors is summarized in Table 2. In the previous studies, if the measurement ranges vary in several tens of degrees or higher, their corresponding sensitivities are only several ‘nm/°C’ or even lower than 1 nm/°C. In contrast, if the sensitivities are ultra-high, e.g. higher than 10 nm/°C, the reported measurement ranges are only several degrees. Hence, the presented fiber optic temperature sensor shows a great performance improvement.
Table 2. Performance comparison of the different fiber optic temperature sensors.
5. Conclusions
In this study, we demonstrate a fiber optic temperature sensor based on PDMS-coated MZI combined with a FBG. It can realize temperature measurement with both high sensitivity and large range. Owing to high RI sensitivity of MZI and high thermo-optic coefficient of PDMS, a high temperature sensitivity can be obtained. Besides, by optimizing the TNCF length, the cascaded FBG can be used to locate different intervals, and therefore the temperature range is greatly extended. The average temperature sensitivities of the sensor can achieve as high as 11.19 nm/°C, 8.53 nm/°C, 7.76 nm/°C, 7.27 nm/°C at the temperature around 30 °C, 40 °C, 50 °C and 60 °C. Besides, the FOMs are as high as 1.92 ${{^{\circ}{\textrm C}}^{ - 1}}$ at 30 °C and 1.37 ${{^{\circ}{\textrm C}}^{ - 1}}$ at 60 °C due to the narrow bandwidth of interference dip. Finally, the proposed sensor shows excellent consistency and repeatability during the continuous thermal cycles, and the maximum temperature deviation rate is only 0.6%.
Funding
National Natural Science Foundation of China (51808347); Natural Science Foundation of SZU (2019109, 860-000002110218).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. R. Min, Z. Liu, L. Pereira, C. Yang, Q. Sui, and C. Marques, “Optical fiber sensing for marine environment and marine structural health monitoring: A review,” Opt. Laser Technol. 140, 107082 (2021). [CrossRef]
2. C. Herrera, G. Nellis, D. Reindl, S. Klein, J. M. Tinjum, and A. McDaniel, “Use of a fiber optic distributed temperature sensing system for thermal response testing of ground-coupled heat exchangers,” Geothermics 71, 331–338 (2018). [CrossRef]
3. Q. Feng, Y. Liang, M. Tang, and J. Ou, “Multi-parameter monitoring for steel pipe structures using monolithic multicore fibre based on spatial-division-multiplex sensing,” Measurement 164, 108121 (2020). [CrossRef]
4. E. Vorathin and Z. M. Hafizi, “Bandwidth modulation and centre wavelength shift of a single FBG for simultaneous water level and temperature sensing,” Measurement 163, 107955 (2020). [CrossRef]
5. M. Chen, Y. Zhao, H. Wei, and S. Krishnaswamy, “Cascaded FPI/LPFG interferometer for high-precision simultaneous measurement of strain and temperature,” Opt. Fiber Technol. 53, 102025 (2019). [CrossRef]
6. H. Li, H. Li, F. Meng, X. Lou, and L. Zhu, “All-fiber MZI sensor based on seven-core fiber and fiber ball symmetrical structure,” Optics and Lasers in Engineering 112, 1–6 (2019). [CrossRef]
7. X. Gao, T. Ning, C. Zhang, J. Xu, J. Zheng, H. Lin, J. Li, L. Pei, and H. You, “A dual-parameter fiber sensor based on few-mode fiber and fiber Bragg grating for strain and temperature sensing,” Opt. Commun. 454, 124441 (2020). [CrossRef]
8. Y. Geng, L. Wang, X. Tan, Y. Xu, X. Hong, and X. Li, “A Compact Four-Wave Mixing-Based Temperature Fiber Sensor With Partially Filled Photonic Crystal Fiber,” IEEE Sens. J. 19(8), 2956–2961 (2019). [CrossRef]
9. M. A. R. Franco, V. A. Serrao, and F. Sircilli, “Side-Polished Microstructured Optical Fiber for Temperature Sensor Application,” IEEE Photonics Technol. Lett. 19(21), 1738–1740 (2007). [CrossRef]
10. N. Zhao, Q. Lin, W. Jing, Z. Jiang, Z. Wu, K. Yao, B. Tian, Z. Zhang, and P. Shi, “High temperature high sensitivity Mach-Zehnder interferometer based on waist-enlarged fiber bitapers,” Sens. Actuators Phys. 267, 491–495 (2017). [CrossRef]
11. Y. Li, L. Wang, Y. Chen, D. Yi, F. Teng, X. Hong, X. Li, Y. Geng, Y. Shi, and D. Luo, “High-performance fiber sensor via Mach-Zehnder interferometer based on immersing exposed-core microstructure fiber in oriented liquid crystals,” Opt. Express 28(3), 3576 (2020). [CrossRef]
12. J. S. Velazquez-Gonzalez, D. Monzon-Hernandez, F. Martinez-Pinon, D. A. May-Arrioja, and I. Hernandez-Romano, “Surface Plasmon Resonance-Based Optical Fiber Embedded in PDMS for Temperature Sensing,” IEEE J. Sel. Top. Quantum Electron. 23(2), 126–131 (2017). [CrossRef]
13. F. Yu, P. Xue, and J. Zheng, “Study of a large lateral core-offset in-line fiber modal interferometer for refractive index sensing,” Opt. Fiber Technol. 47, 107–112 (2019). [CrossRef]
14. J. Zhou and Y. Zheng, “Fiber refractive index sensor with lateral-offset micro-hole fabricated by femtosecond laser,” Optik 185, 1–7 (2019). [CrossRef]
15. C. Lin, Y. Wang, Y. Huang, C. Liao, Z. Bai, M. Hou, Z. Li, and Y. Wang, “Liquid modified photonic crystal fiber for simultaneous temperature and strain measurement,” Photonics Res. 5(2), 129 (2017). [CrossRef]
16. Y. Zhao, Y. Zhang, H. Hu, Y. Yang, M. Lei, and S. Wang, “High-sensitive Mach-Zehnder interferometers based on no-core optical fiber with large lateral offset,” Sensors and Actuators A: Physical 281, 9–14 (2018). [CrossRef]
17. D. Yi, Z. Huo, Y. Geng, X. Li, and X. Hong, “PDMS-coated no-core fiber interferometer with enhanced sensitivity for temperature monitoring applications,” Opt. Fiber Technol. 57, 102185 (2020). [CrossRef]
18. Y. Chen, L. Fang, D. Yi, X. Li, and X. Hong, “Thermo-Optic Property Measurement Using Surface Plasmon Resonance-Based Fiber Optic Sensor,” IEEE Sens. J. 20(19), 11357–11363 (2020). [CrossRef]
19. B.-Y. Ho, H.-P. Su, Y.-P. Tseng, S.-T. Wu, and S.-J. Hwang, “Temperature effects of Mach-Zehnder interferometer using a liquid crystal-filled fiber,” Opt. Express 23(26), 33588–33596 (2015). [CrossRef]
20. K. Tian, G. Farrell, E. Lewis, X. Wang, H. Liang, and P. Wang, “A high sensitivity temperature sensor based on balloon-shaped bent SMF structure with its original polymer coating,” Meas. Sci. Technol. 29(8), 085104 (2018). [CrossRef]
21. R. Fan, L. Li, C. Wang, Z. Du, J. Wang, X. Lv, Z. Ren, and B. Peng, “High-Sensitivity Loop-Fiber Temperature Sensor Based on Distilled Water Cladding,” IEEE Photonics J. 11(5), 1–11 (2019). [CrossRef]
