Open Access
Add to BibSonomy
Abstract
We experimentally demonstrate a highly sensitive directional torsion sensor based on a helical dual side-hole fiber (DSHF). The sensor is constructed by splicing a fraction of DSHF in between two short lengths of multi-mode fibers (MMFs), which forms a Mach-Zehnder interferometer (MZI) based on the interference between the core and cladding modes in the DSHF. The middle part of the DSHF is heated and twisted to get a spiral cladding, so that the cladding mode in the DSHF travels along a helical path. The introduction of the helical structure can greatly improve the torsion sensitivity of the senor and enables it to recognize the twist direction. Experimental results show that the torsion sensitivity can reach as high as 1.666 nm/(rad/m) and −1.413 nm/(rad/m) in the range of 0 ± 5.517 rad/m. Besides, strain sensitivity of the sensor is about −4.5 pm/µɛ.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Optical fiber torsion sensors are of great significance in fields of industry and civil engineering [1,2] due to the advantages of high sensitivity, flexible configuration, compact size and remote sensing ability. A variety of structures have been developed for torsion measurement and the most common are based on gratings [3–7] and interferometers [8–17]. In some practical applications, not only the magnitude of torsion but also the torsion direction should be measured. Therefore, the torsion sensors with direction discrimination ability have attracted much attention. The direction discrimination is usually achieved by introducing a non-circular symmetric structure into the sensing part. Generally, there are two main ways to realize the asymmetry in the fiber. The first one is employing asymmetric fibers such as polarization-maintaining fiber (PMF) or apertured fiber as the sensing unit [8,9]. The second one is to introduce asymmetry structures or to achieve asymmetric refractive index modulation on symmetric fibers. The second way is usually realized by introducing micro machining-‘tapers’ [10–15] or helical structure (HS) in the sensing fiber. These devices taper-based structures have advantages of compact size, easy integration, immunity to electromagnetic interference and high sensitivity. However, most of them may cause mechanical weakness due to the fragile structure made in the tapering process. As for the type with HS, there are mainly two schemes to monitor the torsion. One of the scheme is fabricated by CO2 laser or femtosecond laser. When utilizing the CO2 laser side explosion, two helical or non-circular symmetric perturbations are created in the points of beam splitter and combination [16–18]. By utilizing the femtosecond laser, a helical waveguide (HW) in the cladding of a fiber can be inscribed [19]. The other scheme is fabricated by using a fiber-tapering machine or fusion splicer to make large deformation in the fiber. An interferometer using a single-fiber helix was fabricated by the flame-heated treatment, whose sensitivity has achieved remarkable improvement of 1.691 nm/m/rad [20]. The aforementioned types exhibited good performance, but the CO2 laser induced perturbations can only bring relatively low sensitivity [7], and the fabrication of HW requires ultrahigh precision, which increased production costs and difficulties. In addition, the large deformation of the fibers causes a mechanical weakness.
Recently, some structures with longer HSs and higher torsion sensitivity were proposed and demonstrated. Hailiang Zhang et.al inserted a ∼570 µm helical multicore fiber between two pieces of multi-mode fiber (MMF) [17]. The torsion sensitivity raised to ∼0.118 nm/ (rad/m). Feng Zhang et.al pre-twisted the photonic crystal fibers (PCF) to obtain a good performance [21]. The torsion sensitivity is ∼0.208 nm/ (rad/m) with a twisting period of 1200 µm. The above-mentioned sensors have high sensitivity and strong mechanical endurance, and the manufacturing process is relatively simple.
In this paper, we proposed and experimentally fabricate a highly sensitive torsion sensor based on MZI formed in a dual side-hole fiber (DSHF) with a HS. The proposed structure consists of a fraction of helical DSHF twisted by a fiber-tapering machine and two pieces of MMF. Different from the common torsion sensors, the applying of MZI with HS ensure that the sensor can measure the torsion and determine the twist direction simultaneously. The torsion sensitivity is about one magnitude order larger than the similar structures [17,21]. At last, the sensor has the ability to discriminate axial strain and offers several merits such as repeatability of fabrication, good mechanical robustness structure and compact size, which further benefits its practical sensing applications.
2. Sensor fabrication
The schematic of the proposed MZI is shown in Fig. 1(a). It is composed of a piece of DSHF with a helical structure, which is fusion spliced in between two sections of MMF. The two MMFs are respectively play the role of light splitter and combiner to guide the light from the lead-in SMF into the cladding and core of the DSHF and recouple the light back to the lead-out SMF. The DSHF was fabricated in our laboratory by drilling two air holes in the cladding of a conventional SMF preform and drawing it in a drawing tower with proper air pressure in the air holes. Figure 1(b) shows the cross-section image of the DSHF. The two air holes with diameter of ∼40 µm are located on the two sides of the core. Both the cladding and the core of the DSHF are elliptical. The diameters of the major and minor axes of the elliptical core are ∼10.4 µm and ∼8.7 µm, respectively, and those of the elliptical cladding are ∼127 µm and ∼121 µm, respectively. The side-view microscope photograph of the DSHF with the fabricated HS is shown in Fig. 1(c). From the figure, we can observe that only screw-type deformation is introduced into the DSHF while it is still kept straight as a whole. The length of the HS and the whole DSHF are about 1 cm and 3.3 cm, respectively. The transverse distributions of the optical fields of one cladding mode at the wavelength of 1550 nm in the straight DSHF is shown in Fig. 1(d), and that of the core mode is presented in Fig. 1(e). From the figure, it can be seen that the cladding mode is distributed in a dumbbell shape in the region perpendicular to the line connecting the two air holes. This is because of the existence of the two air hole in the DSHF, which restricted the distribution of the cladding modes.
Fig. 1. (a) Schematic of the proposed MZI. (b) Cross-section image of the DSHF. (c) Side-view microscope photograph of the DSHF with the fabricated HS, (d) the transverse distribution of the optical field for one cladding mode at the wavelength of 1550 nm, and (e) the transverse distributions of the optical fields for core mode.
To fabricate such an integrated structure, firstly, a section of MMF was fusion spliced to a SMF with a single mode fiber splicer (Fujikura S60). The function of the MMF is to couple the light in the lead-in SMF into both the cladding and core of the DSHF as a splitter. Since different modes passing through the MMFs will accumulate phase differences, in order to reduce their influence on the interference spectrum of the sensor, the lengths of the MMFs should be chosen as short as possible to reduce the negative effect of multimode interference on the output spectrum of the sensor. Based on this consideration, the length of the MMFs were chosen to be about 1 mm on the proposed structure. Secondly, the short MMF was spliced to a section of DSHF with a polarization-maintaining fiber fusion splicer (Fujikura FSM-100P) under manual mode. To make sure that the air holes of the DSHF were not collapsed during the process, the arc discharge current and time were set as standard minus 45 bit and 999 ms, respectively. Thirdly, the DSHF was cut off about several centimeters away from the splice junction between the MMF and the DSHF under the monitoring of an optical microscope. Then, the newly cut end surface of the DSHF was cascaded to another MMF with a length of ∼1mm to recouple the two beams in the DSHF to the lead-out SMF with the same operation described in the second step. A MZI was formed due to the effective refractive index difference between the core and cladding modes. At last, the DSHF section of the cascaded structure was heated and twisted with a fiber-tapering machine (Kaipule Co. Ltd. AFBT-8000MX-H), in which way the helical structure was fabricated. The twisted region and helical pitch can be controlled by horizontally movable flame and a pair of rotatable clamps, respectively. The measured transmission spectra of a MZI with a 3.3 cm long DSHF is shown in Fig. 2, the red solid line depicts the spectrum before twist and the black dotted line is the spectrum after twist.
3. Working principle
Since the length of the two MMFs have been chosen to be short enough, the phase difference of their guided modes could be neglected. When the MMF length was 1 mm, almost no interference fringes was observed within the wavelength range of the light source. The light from the lead-in SMF is transmitted into the first MMF which leads the light into the core and cladding of the DSHF, and recoupled to the lead-out SMF by the second MMF, forming the proposed MZI. The transmission of the MZI can be expressed as a two-mode interferometer, and the output intensity of the MZI can be expressed as
where, ${I_{co}}$ and ${I_{cl}}$ are the intensities of the light guided in the core and cladding modes of the DSHF, respectively, and $\varphi = 2\pi ({n_{eff}^{co}{L_{co}} - n_{eff}^{cl}{L_{cl}}} )/\lambda $ is the phase difference between the two light paths, and $\lambda $ is the operational wavelength, $n_{eff}^{co}{L_{co}} - n_{eff}^{cl}{L_{cl}}$ is the phase difference between the core and cladding modes, where $n_{eff}^{co}$ and $n_{eff}^{cl}$ are the effective refractive indices of the core and cladding modes, and ${L_{co}}$ and ${L_{cl}}$ are the physical length of these modes respectively. When the phase difference $\varphi $= (2m + 1) π, m = 0, 1, 2, …., the transmission dips occur at where ${\lambda _m}$ refers to the central wavelength of the mth order interference dip. When the MZI sample is subjected to torsion measurement, the sensitivity of the torsion can be derived as4. Results and discussion
Figure 3 shows the schematic experimental setup for measuring the directional torsion of the sensing element (i.e. the proposed MZI). A broadband light source (SLED) and an optical spectrum analyzer (OSA) were employed. The rotatable clamp has a division value of 5°. The twist rate of the sensing structure can be calculated as $T = \omega /L$, where $\omega $ is the radian of the twist. The sensor is stretched slightly and clamped between two fiber rotators (Thorlabs, PRM1/M) with a distance of ∼28.4 cm. One rotator was rotated in clockwise (CW) or counter clockwise (CCW) direction while the other one was kept fixed. The transmission spectra of the sensor were recorded by varying the rotation angle with a step of 10 degree corresponding to a TR (Technical Record) step of 0.613 rad/m. The TR was increased step by step to −5.517 rad/m and 5.517 rad/m in CCW and CW direction. The TR value was defined as positive in CW direction and negative in CCW direction.
In order to investigate the effects of the pre-twist, we fabricated four samples with different helical pitches and same HS length. Firstly, the response of one sample with DSHF length of 3.3 cm to torsion was measured before it was twisted. The spectrum variation and the wavelength shift versus torsion are presented in Fig. 4, where (a) an (b) are the results of the clockwise and counterclockwise twisting, respectively. The torsion sensitivities in the directions of CW and CCW are 0.042 nm/rad·m−1 and 0.021 nm/rad·m−1, respectively. The result shows that the MZI without HS is extremely insensitive to torsion and has a poor linearity. This is because the external torsion have a small change in the optical path difference (OPD) between the core mode and cladding mode in the MZI.
Fig. 4. Transmission Spectra before the twist under different torsion of the sensor and the relationships between the resonant wavelength and twist rate with the direction of (a) CW and (b) CCW.
Then part of the DSHF in the sample was simultaneously heated and twisted by using a tapering machine to form a permanent HS in the DSHF. The torsion response of the pre-twisted sample was tested again, and the results are shown in Fig. 5. Figures 5(a)–(d) depict the spectral shifts and the corresponding linear approximation with respect to the variation of external torsion under the CW and CCW directions. Four resonant dips (A, B, C and D) in the interferometric spectrum are monitored and the data fitting is performed. From the figures, the spectrum shift to the longer wavelength for CW direction and to the shorter wavelength for CCW direction, which are consistent with the previous theoretical analysis. The corresponding sensitivities are 1.666 nm/rad·m−1, 0.969 nm/rad·m−1, 1.150 nm/rad·m−1, and 0.878 nm/rad·m−1 with the coefficient of determination (R-square) of about 0.997, 0.993, 0.994, 0.997, respectively in CW direction, and those in CCW are −1.337 nm/rad·m−1, −1.410 nm/rad·m−1, −1.413 nm/rad·m−1, −0.767 nm/rad·m−1, with R-square of about 0.969, 0.987, 0.958, 0.978, respectively. From the above results we can get the following points. Firstly the torsion sensitivity of the sample shows a significant improvement, which is ∼40 times higher than the previous MZI without pre-twist and ∼10 times higher than that of the structure in the 7 cores fiber with a helical structure [17]. This is because the solid cladding as well as the air holes at the twisted section are helical, which results in a helical transmission path of the stimulated cladding modes. When an external torsion is applied on the sensing structure, the induced stress will affect the optical path of cladding mode severely, but has little effect on that of the core mode. As a result, the OPD between the core and cladding modes will suffer a great change and causing a large torsion sensitivity. In addition, the torsional sensitivity characterized by the four resonance dips is different. It can be explained that the transmission signal is indeed a multimode interference between the core mode and multiple cladding modes. For a multimode interference, the sensitivity depicted by different dips is usually different, which because different dips correspond to different main cladding modes [22]. At last, the wavelength shifts towards opposite directions for the CW and CCW twist. This is mainly caused by the direction of pre-twist. When the torsion direction is in common with the pre-twist direction, the OPD is increased, which results in a red shift of wavelength accordingly. And the same reason for that of the CCW direction.
Fig. 5. Transmission Spectra after the twist under different torsion of the sensor (a) CW and (c) CCW and the relationships between the resonant wavelength and twist rat with the direction of (b) CW and (d) CCW.
Another sample was fabricated with the same parameters as Sample 2. The new sample has the pitch of 785µm and pitch number of 12, which are similar to those of sample 2. The sensitivities of the new sampler are 1.670 nm/rad·m−1, 1.004 nm/rad·m−1, 1.152 nm/rad·m−1, and 0.883 nm/rad·m−1 in CW direction, and those in CCW are −1.353 nm/rad·m−1, −1.376 nm/rad·m−1, −1.410 nm/rad·m−1, −0.763 nm/rad·m−1, respectively. The sensitivities are also quite similar to those of Sample 2, which means that the structure has a good repeatability.
In order to further investigate influence of the parameters of the helical structure such as helical pitch on torsion sensitivity, two other samples with different pitches were fabricated and measured. All the parameters and results were sorted out in Table 1. The HS’ length of the last 3 samples remains the same, except for the numerical value of the pitch. The torsion sensitivity increase with the increase of the value of pitch in HS, which can also be derived from the previous analysis.
Table 1. Test parameters of 4 samples.
We also measured the temperature and strain effects of the proposed MZI without HS (i.e. sample 1) and with HS (i.e. sample 2), respectively. Temperature response of the proposed device was tested by putting it on a heating furnace with temperature error of ± 1 °C and using two three-dimensional mobile platform to clamp both ends of the sensor to keep it straight. The spectrum responses and wavelength shifts versus temperature of the two samples are shown in Figs. 6(a) and (b), respectively. By linear fitting the measured wavelengths, we can see that the corresponding temperature sensitivities of the sample 1 and sample 2 are respectively 34 pm/°C and 70 pm/°C in the range of 30 °C −100 °C. The highest temperature-torsion cross sensitivity, which is the change of the torsion ratio when the temperature changes by one degree, is 23.8 °C /rad·m−1. It is apparent that the temperature response of the sensor are slightly enhanced after twisting process. This is because the higher cladding mode participating in the interference are gradually away from the interior after twisting process, which make it increasingly sensitive to temperature. But it should be mentioned that both the numerical values are in the same order of magnitude as the common MZIs [14]. The strain response of the proposed device was also tested. The experimental results can be seen from Figs. 7(a) and (b), which shows a sensitivity of −0.6 pm/µɛ with R-square of about 0.979 before the twist and a sensitivity of −4.5 pm/µɛ with R-square of about 0.977 after the twist, respectively. The strain sensitivity of the pre-twisted sample is higher than 7 times of that of the untwisted sample. This is also because a same strain can cause a much larger optical path change of the helical cladding in the pre-twisted DSHF than that in the untwisted DSHF.
Fig. 6. Spectral shifts of the helical DSHF-based MZI at different temperature (a) sample 1 and (b) sample 2, and the corresponding inset show the spectral responses with temperature variation, respectively.
Fig. 7. Spectral shifts of the helical DSHF-based MZI at different strain (a) sample 1 and (b) sample 2, and the corresponding inset show the spectral responses with strain variation, respectively.
5. Conclusion
In conclusion, a highly sensitive directional torsion sensor based on a compact MZI was demonstrated. The MZI was composed of a sandwich structure of SMF-MMF-DSHF-MMF-SMF, in which the DSHF acted as the sensing section and the MMFs served as the beam coupler. A HS structure was formed on the DSHF by heating and twisting the DSHF, which not only improves the torsion sensitivity of the MZI, but also enables it to recognize the direction of twist. Four samples with different pitches were fabricated and tested for comparison. The highest twist sensitivity can respectively reach 1.666 nm/(rad/m) and −1.413 nm/(rad/m) for the CW and CCW twist direction in the torsion range from 0 rad/m to 5.517 rad/m. Besides, we also measured the temperature and strain responses of the sensor, and the sensitivities were 70 pm/°C and −4.5 pm/µɛ, respectively.
Funding
National Defense Pre-Research Foundation of China (6140414030102).
References
1. R. Xu, A. Yurkewich, and R. V. Patel, “Curvature, Torsion, and Force Sensing in Continuum Robots Using Helically Wrapped FBG Sensors,” IEEE Robot. Autom. Lett. 1(2), 1052–1059 (2016). [CrossRef]
2. G. Wang, K. Pran, G. Sagvolden, G. B. Havsgård, A. E. Jensen, G. A. Johnson, and S. T. Vohra, “Ship hull structure monitoring using fibre optic sensors,” Smart Mater. Struct. 10(3), 472–478 (2001). [CrossRef]
3. Z. Han, C. Sun, X. Jin, H. Jinang, C. Yao, S. Zhang, W. Liu, T. Geng, F. Peng, W. Sun, and L. Yuan, “A polarization-independent torsion sensor based on the near-helical long period fiber grating,” Chin. Opt. Lett. 16(10), 100601 (2018).
4. B. Huang and X. Shu, “Ultra-compact strain- and temperature-insensitive torsion sensor based on a line-by-line inscribed phase-shifted FBG,” Opt. Express 24(16), 17670 (2016). [CrossRef]
5. L. Zhang, Y. Liu, B. Cao, and T. Wang, “High Sensitivity Chiral Long-Period Grating Sensors Written in the Twisted Fiber,” IEEE Sens. J. 16(11), 4253–4257 (2016). [CrossRef]
6. L. Xian, P. Wang, and H. Li, “Power-interrogated and simultaneous measurement of temperature and torsion using paired helical long-period fiber gratings with opposite helicities,” Opt. Express 22(17), 20260–20267 (2014). [CrossRef]
7. Y. P. Wang and Y. J. Rao, “Long period fibre grating torsion sensor measuring twist rate and determining twist direction simultaneously,” Electron. Lett. 40(3), 164–166 (2004). [CrossRef]
8. H. M. Kim, T. H. Kim, B. Kim, and Y. Chung, “Temperature-Insensitive Torsion Sensor With Enhanced Sensitivity by Use of a Highly Birefringent Photonic Crystal Fiber,” IEEE Photonics Technol. Lett. 22(20), 1539–1541 (2010). [CrossRef]
9. W. Chen, S. Lou, L. Wang, H. Zou, W. Lu, and S. Jian, “Highly Sensitive Torsion Sensor Based on Sagnac Interferometer Using Side-Leakage Photonic Crystal Fiber,” IEEE Photonics Technol. Lett. 23(21), 1639–1641 (2011). [CrossRef]
10. G. L. Cheng, J. J. Wang, Z. Z. Feng, and N. K. Chen, “Simultaneous twist angle and direction sensing using abrupt-tapered fiber Mach-Zehnder interferometers,” Sixth International Conference on Sensing Technology IEEE, 2013.
11. S. Duan, W. Liu, C. Sun, H. Jiang, C. Yao, K. Zhang, X. Bai, W. Wang, C. Lu, T. Geng, F. Peng, and L. Yuan, “High Sensitive Directional Twist Sensor Based on a Mach–Zehnder Interferometer,” IEEE Photonics J. 10(6), 1–7 (2018). [CrossRef]
12. M. Bianchetti, M. S. Avila-Garcia, R. I. Mata-Chavez, J. M. Sierra-Hernandez, L. A. Zendejas-Andrade, D. Jauregui-Vazquez, J. M. Estudillo-Ayala, and R. Rojas-Laguna, “Symmetric and Asymmetric Core-Offset Mach-Zehnder Interferometer Torsion Sensors,” IEEE Photonics Technol. Lett. 29(18), 1521–1524 (2017). [CrossRef]
13. H. M. Eduardo, S. H. Juan, M. C. Ruth, J. V. Daniel, C. G. Arturo, E. A. Julian, and R. M. Roberto, “A Core-Offset Mach Zehnder Interferometer Based on A Non-Zero Dispersion-Shifted Fiber and Its Torsion Sensing Application,” Sensors 16(6), 856 (2016). [CrossRef]
14. N. K. Chen, J. J. Wang, G. L. Cheng, W. H. Cheng, and P. P Shum, “Fiber torsion sensor with directional discrimination based on twist-induced circular birefringence in unbalanced Mach-Zehnder interferometer,” Lasers & Electro-optics IEEE, 2014.
15. Q. Zhou, W. Zhang, L. Chen, T. Yan, and B. Wang, “Fiber torsion sensor based on a twist taper in polarization-maintaining fiber,” Opt. Express 23(18), 23877–23886 (2015). [CrossRef]
16. F. Qinjiang, Z. Jingdong, L. Chuancan, I. I. Paul, Y. Guolu, L. Lei, S. Yifan, L. Liu, L. Dongxiao, and Z. Tao, “Intensity-modulated directional torsion sensor based on in-line optical fiber Mach–Zehnder interferometer,” Opt. Lett. 43(10), 2414 (2018). [CrossRef]
17. Z. Hailiang, W. Zhifang, S. P. Ping, S. Xuguang, W. Ruoxu, D. X. Quyen, F. Songnian, T. Weijun, and T. Ming, “Directional torsion and temperature discrimination based on a multicore fiber with a helical structure,” Opt. Express 26(1), 544–551 (2018). [CrossRef]
18. Z. Bai, M. Deng, S. Liu, Z. Zhang, J. S. Xu, J. Tang, Y. Wang, C. Liao, and Y. P. Wang, “Torsion Sensor with Rotation Direction Discrimination Based on a Pre-twisted In-Fiber Mach–Zehnder Interferometer,” IEEE Photonics J. 9(3), 1–8 (2017). [CrossRef]
19. L. A. Fernandes, J. R. Grenier, J. S. Aitchison, and P. R. Herman, “Fiber optic stress-independent helical torsion sensor,” Opt. Lett. 40(4), 657–660 (2015). [CrossRef]
20. X.-Y. Li, W.-G. Zhang, L. Chen, Y.-X. Zhang, S. Wang, T.-Y. Yan, Q. Zhou, and B. Wang, “Helical Fiber Interferometer Using Flame-Heated Treatment for Torsion Sensing Application,” IEEE Photonics Technol. Lett. 29(1), 161–164 (2017). [CrossRef]
21. F. Zhang, S. Liu, Y. Wang, Y. Huang, X. Xu, C. Fu, T. Wu, C. Liao, and Y. Wang, “Highly sensitive torsion sensor based on directional coupling in twisted photonic crystal fiber,” Appl. Phys. Express 11(4), 042501 (2018). [CrossRef]
22. Y. Zhang, A. Zhou, and Q. Xu, “Refractive index sensing characteristics of single-mode fiber-based modal interferometers,” J. Lightwave Technol. 32(9), 1734–1740 (2014). [CrossRef]
