Open Access
Add to BibSonomy
Abstract
Combined with non-adiabatic tapering and mode field mismatch between two different fibers, a splicing point tapered fiber Mach-Zehnder interferometer is proposed. Theoretically, two sensing units are carefully designed for dual-parameter sensing, and mode exciting and evolution in fibers are analyzed and shown dynamically. In experiment, transmission spectrum is obtained with two sets of clear interference as designed. After a simple capsulation, simultaneous measurement of seawater temperature and salinity are realized with sensitivities of −994.83pm/°C and 290.47pm/‰, respectively. Additionally, short response time of 33ms and good repeatability are also demonstrated. And effects of encapsulation on avoiding strain and press are verified experimentally. The MZI demonstrated here shows advantages of low cost, easy fabrication, simple construction, compact and robust structure, and dual-parameter sensing with high sensitivity and fast response.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Salinity is an essential quantity to evaluate many marine physical phenomena and dynamical process in ocean, but also a quantity hard to define due to its complex biochemical composition [1]. At present, the most commonly used salinity in marine investigation is practical salinity, which is usually measured with conductivity cells, such as conductivity, temperature, and depth sensor system (CTD). Due to the working principle, this system depends on the constant composition of standard seawater. However, many researches have shown the difference between practical salinity and absolute salinity, especially, the difference is emphasized when the seawater composition is far from that of standard seawater sample. In other words, the standard seawater sample based method, such as CTD, cannot obtain accurate absolute salinity when composition of seawater is not constant. More importantly, instead of practical salinity, absolute salinity is the physical quantity that should be used as official description of seawater in marine science and a key parameter in accessing the thermodynamic properties of the ocean [2], which indicates the importance of measuring absolute salinity accurately. However, the absolute salinity measurement used in marine investigation is still in the exploration stage. To assess the absolute salinity of seawater with optical method, one of the most important methods is refractive index (RI) measurement [1]. However, it is well known that the RI in seawater is both the function of temperature and salinity. In other words, to obtain the absolute salinity accurately, temperature compensation or simultaneous measurement of temperature is required.
Recent years, many optical sensors fabricated by optical fibers have been developed for simultaneous measurement of temperature and salinity in seawater, such as Fiber Bragg grating (FBG) [3–7], fiber Fabry-Pérot interferometer (FPI) [8], microfiber directional coupler [9], and microfiber Mach-Zehnder interferometer (MZI) with a knot resonator [10], show advantages of small size, absolute salinity measured, and immunity to external electromagnetic interference than those of traditional electric method, such as CTD. However, most of the above mentioned sensors need high cost equipment or careful and complex manufacturing process. As a commonly used interference device, all fiber MZIs show prominent advantages of simple fabrication, low cost and compact size. Especially, in line MZIs based on one or two tapered waists have been widely used in filter [11,12], hydrophone [13], pulse shaping [14], optical manipulation [15] and many sensing applications, such as curvature [16], magnetic field [17], gas or liquid refractive index [18–26], and nitrate concentration [27]. During the fabrication of these devices, only one or two tapering steps are required. However, due to the lack of careful design on the spectrum or sensing performance, many sensors suffer from problems like for only single parameter sensing, low sensitivity, or too small extinction ratio observed (around 2-3 dB). In addition, during some sensor fabrications, photonic crystal fiber (PCF) is used and shows advantages of simple structure, specificity detection, and low temperature sensitivity, however, for a temperature sensing application, low temperature sensitivity is not desired.
In this paper, based on cheap and commonly used optical fibers, such as thin-core fiber (TCF) and standard single-mode fiber (SMF), a splicing point tapered fiber MZI for simultaneous measurement of seawater temperature and salinity is proposed, in which TCF is designed for temperature compensation and tapered splicing points between TCF and SMF help to excite stronger high order mode, which is beneficial for higher extinction ratio and sensing sensitivity. To realize the above conception, structure of MZI (including the splicing method, length and type of TCF) is designed using the beam propagation theory. Especially, to show the mode excitation process more clearly, relative power of different modes excited at different positions is shown by a dynamic graph. Based on the theoretical analysis and design, corresponding structure is fabricated and transmission spectrum is obtained with two sets of clear interference for dual-parameter sensing as expected. To make the tapered splicing point more robust and keep the sensing function in the sensitive area, a practical and simple half encapsulation method is introduced, by which cross-sensitivity between strain and curvature can also be avoided. Using the fabricated MZI, simultaneous measurement of seawater temperature and salinity are realized, and sensitivities are estimated to be −777.90pm/°C and 223.07pm/‰, respectively. By further analysis on sensitivity, sensitivities of −994.83pm/°C and 290.47pm/‰ can be obtained, which are more than 100 times higher than those of FBGs, FPIs, and multimode fibers in temperature sensing and 2-23 times higher than those of FBG, PCF, microfiber knot resonator, and two-core fiber in salinity sensing. In addition, response time and repeatability are also investigated. Results show its good repeatability and fast response time of about 33ms. In addition, effects of encapsulation on avoiding strain and press are also demonstrated experimentally. MZI demonstrated here shows advantages of simple and compact construction, robust structure, easy fabrication and packaging, dual-parameter sensing, high sensitivity, fast response, and good repeatability. And the methods of half encapsulation and designing a section of TCF as temperature compensation or dual-parameter sensing may afford useful references for other sensing applications, such as pressure sensing, curvature sensing, gas refractometers and so on.
2. Design and fabrication
Generally, for interference devices, to realize dual-parameter sensing, such as simultaneous measurement of temperature and salinity in seawater, the most direct way is to establish an interference structure with two sets of clear interference. It is well known that both non-adiabatic tapering of a fiber and the mode field mismatch between two different fibers can excite the high order mode or cladding mode, and the interference between the core mode (fundamental mode) and cladding mode (high order mode) results in the interference spectrum we observed, which is the main working principle for almost all in-line MZIs. Combined with non-adiabatic tapering and mode field mismatch between two different fibers, we design a splicing point tapered fiber MZI assembled by two different fibers to produce such a spectrum with two sets of interference required.
2.1 Design on the first modal interference unit
The sensing structure designed is shown in Fig. 1(a) based on the following considerations: Firstly, considering the low cost, we choose the SMF (SMF-28e) produced by Corning Co. and TCF (Product Number 1060-XP, with a core diameter of 5.8μm and price of $2.45/meter) produced by Nufern Co. as two different fibers to assemble this sensor. It's easy to imagine that for a spectrum containing interferences from both non-adiabatic tapering and mode field mismatch, two interferences will disturb each other easily, if without careful design.
Fig. 1 (a) Schematic of the proposed MZI structure. (b) Different structures during the manufacturing process. (c) Contrasting our result with other result. Calculated relative power of LP0n mode, LP1n mode and LP2n mode for (d) structure A, (e) B, and (f) C, respectively.
To reveal the effects of non-adiabatic tapering and the mode field mismatch between TCF and SMF, respectively, different structures (A, B, C, and D) during the manufacturing process are plotted in Fig. 1(b), and relative power of LP0n mode, LP1n mode and LP2n mode for each structures are analyzed based on the beam propagation theory, which can be realized by a commercial software “Optiwave” [28,29]. To verify the correctness of our modeling with the software, Tian’s result published was firstly repeated by establishing the same modeling [28]. Our result shows the good agreement with Tian’s simulations, as shown in Fig. 1(c). The little difference between them is mainly caused by the different version software used.
Calculated relative power of LP0n mode, LP1n mode and LP2n mode for structure A, B and C under 1550-nm-wavelenght are shown in Figs. 1(d)-1(f), respectively. For structure A, 5-cm-length 1060-XP TCF is spliced between two SMFs without any offset. Similar with other demonstrated in-line MZIs, at the splicing joint light propagating in the lead-in SMF (SMF1) core is divided into two parts: a fraction of light propagates into the core of TCF as the core mode and a fraction of light propagates into the cladding of TCF as the cladding mode. At the second splicing point, the cladding modes interfere with the core mode due to the phase difference, thus the TCF acts as the first inter-modal interference sensing unit. However, due that the core diameter of TCF is very closely to that of SMF, most energy exist in the mode of LP01 mode (fundamental mode) and little cladding mode is excited, which can be clearly seen from Fig. 1(d). Considering that tapering process can also help to excite cladding mode, to excite more cladding mode, the first splicing point between SMF1 and TCF is tapered like structure B and the detail of the taper is shown in the inset of Fig. 1(e). As a result, more fundamental mode is coupled into higher order mode with the same azimuthal symmetry LP0n mode since the tapers are axisymmetric, as shown in Fig. 1(e). It can be expected that more cladding mode can be furtherly excited by tapering the second splicing point between the TCF and SMF2, like structure C. As shown in Fig. 1(f), more fundamental mode couples into high order mode. As a result, energy of high order mode becomes to be much closer to that of fundamental mode, which indicates that higher extinction ratio of spectrum can be obtained. Based on the above analysis, we can conclude that the section of TCF is designed to act as the first interference sensing unit with relatively high extinction ratio due to the strong cladding mode excited by two tapers at the splicing point between TCF and SMF.
2.2 Design on the second modal interference unit
Secondly, as demonstrated in recent work shown in reference 24 and 27, when light is launched into a tapered SMF in the form of fundamental mode, high order mode may be excited in the tapered transition. It must be pointed out that the high order mode cannot be always excited unless the transition meets the non-adiabatic condition. To ensure the non-adiabatic condition, the second tapering point between TCF and SMF2 is further tapered to be the structure D and the second modal interference sensing unit can be obtained.
To verify the above analysis, structures A, B, C and D are fabricated successively in experiment. Fabrication steps are described in Figs. 2(a)-2(d). Structure A is fabricated by splicing SMF and TCF directly. Based on the fabrication of structure A, structure B and C can be obtained by tapering the splicing point between SMF and TCF under the arc-discharge of the fusion splicer. Finally, structure D is completed by further tapering the second splicing point under the alcohol lamp heating. The reason why arc-discharge of welder firstly used and then alcohol lamp used is to ensure the non-adiabatic tapering condition, and make the device more compact and easier to encapsulate. The optical micrographs of the completed taper A and taper B are shown in Figs. 2(e) and 2(f), respectively. It can be measured that for taper A and B, lengths of waist region are about 26.9μm and 1308.8μm, diameters of waist region are about 33.2μm and 9.5μm, and lengths of transition region are about 252.0μm and 313.7μm, respectively.
Figures 2 (a) - 2(c) Fabrication steps of taper A or B. (d) Further tapering on taper B by alcohol lamp heating. Optical micrographs of the completed (e) taper A and (f) taper B.
During the fabrications of the above structures, transmission spectra of every structure are also monitored simultaneously, which are shown in Fig. 3(a). It can be seen that from structure A to structure C the extinction ratio becomes to be higher and higher due to more and more cladding mode excited, which is consistent with above analysis. Compared the spectra of structure C and D, it can be found that the spectra of structure D can be considered to be an envelope modulated spectra based on structure C, and the envelope comes from the further tapering of the second splicing point under the alcohol lamp heating.
Fig. 3 (a) Transmission spectra of structures during the fabrications. (b) The Fourier transform of the spectra of the structure C, and D before and after FFT smooth.
To further demonstrate it, a low pass filtering algorithm based on fast Fourier transform (FFT) smoothing method is firstly used, by which the curve is smoothed by filtering out the high frequency signals. The smoothed curve of the black line (structure D) is shown in dark blue in Fig. 3(a). Then by Fourier transform of the black, blue and pink lines in Fig. 3(a), we can get their corresponding spatial frequency spectra, as shown in Fig. 3(b). It can be seen from the black line that there are only one dominantly excited high order mode, along with other weakly excited ones. Compared the spatial frequency spectra before (black line in Fig. 3(b)) and after smoothing (dark blue line in Fig. 3(b)), we can conclude that the smoothing process can really filter the high frequency signals. In addition, compared with the spatial frequency spectra before (pink line in Fig. 3(b)) and after flame tapering the second splicing point (black line in Fig. 3(b)), we can find that the dominant excited mode is mainly derived from the flame tapering process.
Based on the above results, we can conclude that both sides tapering on the splicing point with fusion splicer excite the higher order modes. Further tapering of the second splicing point under the alcohol lamp will excite more high order mode and produce a dominantly excited mode. The interference between the high order mode and fundamental mode happened in section of thin-core fiber act as the first modal interference sensing unit. And the interference between the dominantly excited mode and fundamental mode happened in taper B can be used as the second modal interference sensing unit.
2.3 Mode exciting and evolution process
To reveal the mode exciting and evolution process more clearly, dependences of relative power of each mode on different positions are calculated. As shown in Figs. 1(d)-1(f), little LP1n and LP2n mode is excited, so we only calculate the LP0n mode. Take the similar structure used in Fig. 3(a) for example, the calculated mode exciting and evolution process during the propagation is shown in Visualization 1. Light with 1550-nm-wavelength is launched into the SMF1 in LP01 mode. When light arrives at the transition region of SMF1-TCF, power of LP01 mode decreases gradually, accompanying with the increasing power of other high order modes, as shown in Fig. 4(c). Additionally, it can be seen that mode exciting mainly occurs in the SMF1-TCF down-taper region and keeps the exciting process until the up-taper region.
Fig. 4 (a) Movie of the mode exciting process in MZI with typical wavelength of 1550 nm (see Visualization 1). Enlarged details of (b) taper A and (c) taper B in Fig. 4(a). (d) Movie of the mode exciting process in MZI with typical wavelength of 1550 nm when section of TCF is changed into SMF with the same length (see Visualization 2). Enlarged details of (e) taper A and (f) taper B in Fig. 4(d).
To estimate the individual contribution of the non-adiabatic tapering and the mode field mismatch between two different fibers in mode exciting, we replace the section of TCF with SMF and other parameters keep unchanged. The obtained mode exciting and coupling process are shown in Visualization 2, and tapered regions are enlarged in Fig. 4(e) and 4(f), respectively. By contrasting Figs. 4(b) and 4(e) or Figs. 4(c) and 4(f), it can be found that the mode field mismatch between two different fibers can really help to excite stronger high order mode. In addition, the energy oscillation of LP01 mode and other high order mode in waist region shows the interference clearly.
2.4 Design on the type of TCF
Finally, it is noticing that for this structure the section of thin-core fiber 1060-XP cannot be replaced by other thin-core fibers, such as 780-HP (with a core diameter of 3.6μm) and 980-HP (with a core diameter of 4.4μm), because the core diameter difference between the 780-HP (or 980-HP fiber) and SMF are much larger than that of 1060-XP fiber. The larger core diameter difference will lead to much stronger excitation of high order mode and more significant interference of the first modal interference sensing unit, and thus it will cover up or mix with the interference of the second modal interference sensing unit. Figure 5(a) shows the transmission spectra of structure A, B, C, D when the section of 1060-XP replaced by the 780-HP fiber and all other parameters remain the same. As is shown, the interferences of structure B and C show much higher extinction ratio than those shown in Fig. 3(a), and the purple line indicates the mixture of two sets of interference. To see more clearly, FFT smooth process is also performed on the purple line, and result is shown in blue in Fig. 5(a).
Fig. 5 (a) Transmission spectra of the structures during fabrications when the section of 1060-XP is replaced by the 780-HP fiber or 980-HP fiber. Calculated relative power of LP0n mode, LP1n mode and LP2n mode for structure B when (b) 780-HP fiber used or (c) 980-HP fiber used.
To further verify the stronger excitation of high order mode when 780-HP fiber is used, we calculate the mode energy of fundamental mode and high order mode in structure B. As is shown in Fig. 5(b), high order mode has almost the equal energy with fundamental mode when 780-HP TCF used, and the equal energy between them will generate an interference spectra with large extinction ratio. For contrast, when 1060-TCF is used, as shown in Fig. 1(e), energy difference between fundamental mode and high order mode is relatively large.
Besides the 780-HP fiber, when the section of 1060-XP is replaced by the 980-HP fiber and all other parameters remain the same, similar experimental and theoretical results are also obtained, as shown in Figs. 5(a) and 5(c), respectively.
On the other hand, TCF with core diameter much larger than that of 1060-XP is also not adopted due to the insufficient excitation of high order mode, which will lead to the weak interference between fundamental mode and high order mode happened in the first interference unit and low extinction ratio of the sensing peak. Though stronger high order mode can also be excited by decreasing the diameter of taper A and taper B, however, this will weaken the robustness of the structure.
3. Results of sensing experiments
3.1 Design on the device encapsulation
Based on the above structure design and fabrication, sensing experiment can be performed. Due to the redrawing of the second splicing point, it is necessary to make this tapered splicing point more robust. Additionally, to keep the sensing function of the sensitive area, a half encapsulation method is introduced as follows: firstly, moderate liquid Polydimethylsiloxane (PDMS) is injected into the bottom of a partial removed metal tube. By heating metal tube under 120°C for 15 minutes, PDMS solidified. Then put the taper B into the metal tube. Drop little liquid PDMS at the untapered region, and solidified it with the same method. The schematic diagram and optical micrograph of the sealing structure is shown in Fig. 6(a). To evaluate the influence of PDMS, transmission spectra before and after sealing are contrasted in Fig. 6(b). It can be seen that the spectrum can still keep clear interference after sealing.
Fig. 6 (a) Schematic of the capsulation method and the photo of MZI after sealing. Transmission spectra of MZI (b) before and after sealing and (c) before and after immersing into water. Inset: the Fourier transform of the spectra of the fabricated structure D in water before and after FFT smooth.
After sealing the structure in air, the whole structure is then immersed into the distilled water. As shown in Fig. 6(c), due to the change of surrounding material, transmission spectrum also changed. By Fourier transform of the black and blue lines in Fig. 6(c), spatial frequency spectra are obtained and shown in the inset of Fig. 6(c). It can be seen that the frequency of the peak corresponding to the dominantly excited mode remains unchanged before and after the FFT smooth processing, which indicates that we can track the peaks of the smoothed line (black line in Fig. 6(c)) to reveal the response of the second sensing unit on the change of salinity or temperature.
3.2 Salinity and temperature sensing experiments
With the sealed structure, salinity sensing is firstly performed with the sensing system shown in Fig. 7(a), which consists of a broad-band light source (SuperKTM Compact), a fabricated sensor, an optical spectrum analyzer (OSA, Ando AQ6379C), a fiber demodulator, a temperature-controlled heating platform, a thermocouple thermometer (TCT) and salinity meter (Salimeter). It is noticing that to avoid the inconvenience in actual signal acquisition and processing with OSA, and to acquire in time signals of peak A and B for real-time detection, a fiber demodulator can be introduced in the sensing system, as shown in Fig. 7(a). In such a system, OSA is used for collecting and analyzing the original signal of peak B; Demodulator is used for completing the filtering algorithm process and then outputting the signal of peak A.
Fig. 7 (a) Schematic of sensing system. Transmission spectra during salinity sensing experiment (b) without and (c) with FFT smoothed. (d) Peak B shifts with the increasing salinity; Insets in Figs. 7(c) and 7(d): linear fittings of the wavelength of sensing peak A and B at different salinities.
To demonstrate the salinity sensing, high salinity seawater is added into the distilled water in successive and the salinity of sample is artificially tuned from 0‰ to 49‰. The original transmission spectra under different salinities are recorded in Fig. 7(b). For convenience of signal analysis, high frequency signals are filtered from the original transmission spectra with the above used low pass filtering algorithm and the obtained spectra has been shown in black line of Fig. 6(c).
By tracking one of the typical peaks, such as peak A, we plot the peak shifts with the increasing salinities in the inset of Fig. 7(c). It can be seen that the peak shifts to the long wave band when salinity increases with sensitivity of 223.07pm/‰. In addition, it can be predicted that due to the large diameter of taper A and 1060-XP fiber, very little evanescent field is excited in the first inter-modal interference sensing unit and thus salinity sensitivity should be negligible. To verify it, a typical peak, such as peak B, is selected and by tracking the shift of peak B, salinity sensitivity of the first inter-modal interference sensing unit is estimated to be about 5.7pm/‰, as shown in Fig. 7(d).
Then temperature sensing is performed as follows: by controlling the temperature of heating platform, temperature of seawater is tuned from 18.9°C to 39.6°C, and TCT is used for calibration. Similarly, by tracking peak A and B, sensitivities of temperature sensing of peak A and B are −777.90pm/°C and 89.81pm/°C, respectively, as shown in Figs. 8(a) and 8(b).
Fig. 8 (a) FFT smoothed transmission spectra during temperature sensing experiment. (b) Peak B shifts with the increasing temperature; Insets in Figs. 8(a) and 8(b): Linear fittings of the wavelength of the sensing peak A and B at different temperatures.
3.3 Calibration and sample test
Based on the above measurement, a sensitivity matrix can be established as follows:
Considering the good linearity of the above fittings, seawater with temperature of 22.3°C and salinity of 30‰ are selected as the standard sample to calibrate the system. By obtaining the wavelengths of the peak A of 1405.36nm and peak B of 1445.56nm under the calibration temperature and salinity, simultaneous measurements of temperature T and salinity S can be realized by tracking peak A λA and peak B λB at any time using the following relationship:
To verify the feasibility and accuracy of the above mentioned calibration method, we test two samples under random temperature and salinity. By scanning the transmission spectra of two samples, as shown in Fig. 9(a) and tracking the wavelength of peak A and peak B, as shown in Figs. 9(b) and 9(c), temperatures and salinities of the samples can be obtained using the Eq. (2). Table 1 shows the comparisons of the results we measured with those measured by TCT and salimeter. It can be seen that results we measured show good agreement with those commonly used commercial devices measured.
Fig. 9 (a) Transmission spectra of standard sample, test 1, and test 2 sample. Tracking the (b) peak A and (c) peak B.
Table 1. Comparisons of two tests on temperatures and salinities with commercial devices
4. Discussions
4.1 Effect of peak wavelength on sensitivity
Generally speaking, for most interferences in MZI, with the increasing wavelength, sensitivity usually increases gradually, which is true for these peaks coming from the second modal interference unit. Figures 10(a) and 10(b) show the positive relations between the sensitivities and probing wavelength clearly. However, for those peaks coming from the first modal interference unit, things will be different. It can be seen from Fig. 10(c) that with the increasing wavelength, sensitivity for temperature sensing firstly goes up and then goes down, which dues to the opposite temperature response of the two sensing units. Because of the low salinity sensitivity of the first interference unit, dependence of salinity sensitivity on wavelength is not discussed.
Fig. 10 Dependences of (a) salinity and (b) temperature sensing sensitivities on dip wavelength in FFT smoothed spectra. (c) Dependence of temperature sensing sensitivity on dip wavelength in unsmoothed spectra.
4.2 Response time of sensor
For further in situ and real-time sensing, especially, the process happening on a time scale of microseconds, such as turbulence in ocean, the response time of the sensor should be concerned. Considering that both temperature and salinity diffusion in liquid will take amount of time, and the time mainly depends on the heat or solute diffusion rate, which will effect the response time of sensor measured inevitably. So response time of the sensor is measured by throwing the sensor from air into seawater directly to reveal the real response of the sensor to the change of environment because both the change of temperature and salinity will lead to the change of refractive index of seawater. The obtained temporal response of the output under 1485nm is given in Fig. 11(a). Considering that in fiber optic communications, response time is usually defined as the time reaching 90% of the variation of the surroundings. Thus, the response time of sensor can be estimated to be about 33ms, which can satisfy the need of turbulence measurement in ocean basically.
Fig. 11 (a) Temporal response of the output under 1485nm-wavelength. Inset: the enlarged detail of the response. (b) Repeatability test of the sensor. Inset: the original transmission spectrum under 13.2°C and peak A tracked.
4.3 Repeatability
Besides the temporal response, repeatability of the sensor is also tested. Take the temperature sensing for example, firstly, an original transmission spectrum under 13.2°C is collected and a typical peak, such as the peak A around 1463nm is chosen to be tracked. Then during a temperature-up from 13.2°C to 24.6°C and temperature-down process, positions of this peak under typical temperatures are all recorded. Repeating the above process three times, finally, a complete picture describing the repeatability of the sensor is obtained. Figure 11(b) shows the dependences of peak A positions on temperatures. It seems that for all tests the linearly fittings lines of temperature-up process and down processes do not fit well with each other. However, this mainly dues that during the temperature sensing process, because of the small quantity of seawater, evaporation induced salinity changing is noticeable, and the changing salinity also contributes the peak shift. As the experiment goes on, salinity of sample becomes to be higher and higher (from 32‰ to 37‰), thus from test 1 to test 3, the sensing peak keeps going towards the long wave direction. Especially, compared with temperature-down process, temperature-up process always shows more violent evaporation of seawater and faster salinity changing, as a result, the sensitivity of temperature-up process shows lower sensitivity than that of temperature-down process due to the opposite temperature and salinity response of this sensor. Even so, the repeatability can also be evaluated by just comparing the heat up or cool down process in different tests. The almost identical slope of all temperature-up or down lines indicates the good repeatability of the sensor. Because in real seawater, the amount of sea water is very large and the change of salinity induced by liquid evaporation is negligible.
4.4 Effects of encapsulation on avoiding strain and press
Besides protecting the taper B from broken, the encapsulation can also avoid the effects of strain and press. To demonstrate that the encapsulation can really avoid the effect of strain, experimental setup shown in Fig. 12(a) is established by referring to the published work [30]. As is shown, the tested sensor is clamped between a translation stage and a fixed stage. By moving the translation stage, strain is applied on the tested sensor from 0 to 300 or 600με.
Fig. 12 (a) Schematic of strain test system. (b) Transmission spectra of sensor without encapsulation under different strains. (c) FFT smoothed spectra of sensor without encapsulation under different strains. (d) Transmission spectra of sensor with encapsulated taper B under different strains. (e) FFT smoothed spectra of sensor with encapsulated taper B under different strains. (f) Transmission spectra of sensor with encapsulated taper B under different pressures.
To reveal the effect of the encapsulation on avoiding strain, the sensor without any encapsulation is firstly tested in experiment. As shown in Fig. 12(b), with the increasing strain, transmission spectrum changed gradually. Due that the diameter of taper B is much smaller than that of taper A, the taper B without encapsulation is much more sensitive to the strain. To see the effect of the strain on un-encapsulated taper B, the FFT smooth of the transmission spectrum is also performed and shown in Fig. 12(c). It can be seen that the peak shifts with the increasing strain evidently due to the elastic-optic effect and cavity length change induced by the elongated taper B.
For contrast, the strain experiment is repeated after the encapsulation of taper B on the same sample. The encapsulated structure is shown in the Fig. 12(a). Obtained transmission spectrum is shown in Fig. 12(d) and the signal after FFT smoothing is shown in Fig. 12(e), respectively. It can seen that due to the rigid encapsulating on taper B, the elastic deformation of taper B can be eliminated, thus for the same interference dip, such as the peak pointed with the black arrow in Figs. 12(c) and 12(e), the dip almost keeps unchanged even though larger strain is applied.
Therefore, by comparisons of the above Figs. 12(c) and 12(e), it can be concluded that after encapsulating taper B, the effect of strain on taper B can be avoided successfully. Due the taper A is not sealed, the effect of strain will lead to the extension of the first modal interference sensing unit and the reshape of taper A, thus the peaks coming from this interference unit will shift, especially when the taper B is encapsulated, as shown in Fig. 12(d).
It is noticing that because in our temperature or salinity sensing experiments, no strain is applied, thus the encapsulation is only used in taper B to protect the thin fiber from broken, as shown in Fig. 12(a). However, in strain experiment, when moving the translation stage, both taper A and taper B are affected by the strain. In other words, to ensure the strain insensitivity of the total spectrum, the whole structure of the sensor (including taper A and taper B) should be encapsulated completely, which can be realized by just using a longer metal tube to ensure that the taper A is wrapped inside the tube.
In addition, stress response of the sensor is also investigated, heavy objects with different weights are placed on the encapsulated section of the sensor gradually, and transmission spectra under different weights are shown in the Fig. 12(f). It can be found that the spectrum basically keeps unchanged and indicates the press resistance of the sensor with encapsulation.
Finally, besides the strain and press, when bending force is applied on the metal tube, the rigid metal material will also suffer from this force and protect the fiber inside the half opened metal tube because the sensing unit is free-standing inside the half opened metal tube.
5. Conclusions
In conclusion, combined with non-adiabatic tapering and mode field mismatch between two different fibers, a splicing point tapered fiber MZI for simultaneous measurement of seawater temperature and salinity is proposed based on TCF and SMF. Using the beam propagation theory, relative power of different modes is calculated, and structure of MZI is carefully designed. Especially, to show the mode excitation process more clearly, relative power of different modes excited at different positions is displayed. Based on the theoretical analysis and design, corresponding structure is fabricated and transmission spectrum is obtained with two sets of clear interferences for dual-parameters sensing. To make the tapered splicing point more robust, as well as keep the sensing function of the sensitive area, a half encapsulation method is introduced. Subsequently, simultaneous measurement of seawater temperature and salinity are realized, and sensitivities are estimated to be −777.90pm/°C and 223.07pm/‰, respectively. By choosing larger wavelength in spectra, sensitivities of −994.83pm/°C and 290.47pm/‰ can be obtained, which are more than 100 times higher than those of FBGs, FPIs, and multimode fibers in temperature sensing and 2-23 times higher than those of FBG, PCF, microfiber knot resonator, and two-core fiber in salinity sensing. In addition, short response time of 33ms and good repeatability are also demonstrated. And finally, effects of encapsulation on avoiding strain and press are proved experimentally. The MZI demonstrated here shows advantages of easy fabrication and packaging, dual-parameter sensing, high sensitivity, fast response, and good repeatability. And methods of half encapsulation and designing a section of TCF as temperature compensation or dual-parameter sensing may afford useful references for other sensing applications.
Funding
National Key R&D Program of China (2017YFC1405600); National Natural Science Foundation of China (61871353) and Shandong Provincial Natural Science Foundation, China (ZR2017MD030).
References
1. M. L. Menn, P. A. G. Albo, S. Lago, R. Romeo, and F. Sparasci, “The absolute salinity of seawater and its measurands,” Metrologia 56(1), 015005 (2019). [CrossRef]
2. P. Grosso, M. L. Menn, J.-L. de Bougrenet de la Tocnaye, Z. Y. Wu, and D. Malardé, “Practical versus absolute salinity measurements: New advances in high performance seawater salinity sensors,” Deep Sea Res. Part I Oceanogr. Res. Pap. 57(1), 151–156 (2010). [CrossRef]
3. D. A. Pereira, O. Frazão, and J. L. Santos, “Fiber Bragg grating sensing system for simultaneous measurement of salinity and temperature,” Opt. Eng. 43(2), 299–304 (2004). [CrossRef]
4. L. Q. Men, P. Lu, and Q. Y. Chen, “A multiplexed fiber Bragg grating sensor for simultaneous salinity and temperature measurement,” J. Appl. Phys. 103(5), 053107 (2008). [CrossRef]
5. X. P. Liu, T. T. Wang, Y. Wu, Y. Gong, and Y. J. Rao, “Dual-parameter sensor based on tapered FBG combined with microfiber cavity,” IEEE Photonics Technol. Lett. 26(8), 817–820 (2014). [CrossRef]
6. H. Y. Meng, W. Shen, G. B. Zhang, C. H. Tan, and X. G. Huang, “Fiber Bragg grating-based fiber sensor for simultaneous measurement of refractive index and temperature,” Sens. Actuators B Chem. 150(1), 226–229 (2015). [CrossRef]
7. X. Shu, B. A. Gwandu, Y. Liu, L. Zhang, and I. Bennion, “Sampled fiber Bragg grating for simultaneous refractive-index and temperature measurement,” Opt. Lett. 26(11), 774–776 (2001). [CrossRef] [PubMed]
8. L. V. Nguyen, M. Vasiliev, and K. Alameh, “Three-wave fiber Fabry-Pérot interferometer for simultaneous measurement of temperature and water salinity of seawater,” IEEE Photonics Technol. Lett. 23(7), 450–452 (2011). [CrossRef]
9. S. S. Wang, H. J. Yang, Y. P. Liao, X. Wang, and J. Wang, “High-sensitivity salinity and temperature Sensing in seawater based on a microfiber directional coupler,” IEEE Photonics J. 8(4), 6804209 (2016). [CrossRef]
10. Y. P. Liao, J. Wang, S. S. Wang, H. J. Yang, and X. Wang, “Simultaneous measurement of seawater temperature and salinity based on microfiber MZ interferometer with a knot resonator,” J. Lightwave Technol. 34(23), 5378–5384 (2016). [CrossRef]
11. X. Wang, Y. Li, and X. Bao, “C- and L-band tunable fiber ring laser using a two-taper Mach-Zehnder interferometer filter,” Opt. Lett. 35(20), 3354–3356 (2010). [CrossRef] [PubMed]
12. G. S. Delgado, A. M. Rios, and D. M. Hernandez, “Tailoring Mach-Zender comb filters based on concatenated tapers,” J. Lightwave Technol. 31(5), 761–767 (2013). [CrossRef]
13. S. Dass and R. Jha, “Cascaded Taper Mach-Zehnder Interferometer Based Hydrophone,” in 13th International Conference on Fiber Optics and Photonics (Optical Society of America, 2016), paper W4G.4.
14. S. Mas, J. Palací, P. P. Millán, S. Lechago, D. M. Hernández, and J. Martí, “All-fiber processing of terahertz-bandwidth signals based on cascaded tapered fiber,” Opt. Lett. 38(23), 4954–4957 (2013). [CrossRef] [PubMed]
15. Q. Rong, Y. Zhou, X. Yin, Z. Shao, and X. Qiao, “Higher-order micro-fiber modes for Escherichia coli manipulation using a tapered seven-core fiber,” Biomed. Opt. Express 8(9), 4096–4107 (2017). [CrossRef] [PubMed]
16. S. Dass and R. Jha, “Micrometer wire assisted inline Mach-Zehnder interferometric curvature sensor,” IEEE Photonics Technol. Lett. 28(1), 31–34 (2016). [CrossRef]
17. B. Sun, F. Fang, Z. Zhang, J. Xu, and L. Zhang, “High-sensitivity and low-temperature magnetic field sensor based on tapered two-mode fiber interference,” Opt. Lett. 43(6), 1311–1314 (2018). [CrossRef] [PubMed]
18. D. Wu, T. Zhu, M. Deng, D. W. Duan, L. L. Shi, J. Yao, and Y. J. Rao, “Refractive index sensing based on Mach-Zehnder interferometer formed by three cascaded single-mode fiber tapers,” Appl. Opt. 50(11), 1548–1553 (2011). [CrossRef] [PubMed]
19. J. Yang, L. Jiang, S. Wang, B. Li, M. Wang, H. Xiao, Y. Lu, and H. Tsai, “High sensitivity of taper-based Mach-Zehnder interferometer embedded in a thinned optical fiber for refractive index sensing,” Appl. Opt. 50(28), 5503–5507 (2011). [CrossRef] [PubMed]
20. D. Wu, Y. Zhao, and J. Li, “PCF taper-based Mach-Zehnder interferometer for refractive index sensing in a PDMS detection cell,” Sens. Actuators B Chem. 213, 1–4 (2015). [CrossRef]
21. Q. Wang, W. Q. Wei, M. J. Guo, and Y. Zhao, “Optimization of cascaded fiber tapered Mach-Zehnder interferometer and refractive index sensing technology,” Sens. Actuators B Chem. 222, 159–165 (2016). [CrossRef]
22. Y. Zhao, F. Xia, and J. Li, “Sensitivity-enhanced photonic crystal fiber refractive index sensor with two waist-broadened tapers,” J. Lightwave Technol. 34(4), 1373–1379 (2016). [CrossRef]
23. Q. Wang, L. X. Kong, Y.L. Dang, F. Xia, Y. W. Zhang, Y. Zhao, H. F. Hu, and J. Li, “High sensitivity refractive index sensor based on splicing points tapered SMF-PCF-SMF structure Mach-Zehnder mode interferometer,” Sens. Actuators B Chem. 225, 213–220 (2016). [CrossRef]
24. N. M. Y. Zhang, K. Li, N. Zhang, Y. Zheng, T. Zhang, M. Qi, P. Shum, and L. Wei, “Highly sensitive gas refractometers based on optical microfiber modal interferometers operating at dispersion turning point,” Opt. Express 26(22), 29148–29158 (2018). [CrossRef] [PubMed]
25. C. Tao, H. Wei, and W. Feng, “Photonic crystal fiber in-line Mach-Zehnder interferometer for explosive detection,” Opt. Express 24(3), 2806–2817 (2016). [CrossRef] [PubMed]
26. X. Feng, W. L. Feng, C. Y. Tao, D. S. Deng, X. Qin, and R. Chen, “Hydrogen sulfide gas sensor based on graphene-coated tapered photonic crystal fiber interferometer,” Sens. Actuators B Chem. 247, 540–545 (2017). [CrossRef]
27. J. Wang, Y. Liao, S. Wang, and X. Wang, “Ultrasensitive optical sensing in aqueous solution based on microfiber modal interferometer,” Opt. Express 26(19), 24843–24853 (2018). [CrossRef] [PubMed]
28. Z. B. Tian and S. S.-H. Yam, “In-line single-mode optical fiber interferometric refractive index sensors,” J. Lightwave Technol. 27(13), 2296–2306 (2009). [CrossRef]
29. “Optiwave,” Opti BPM [Online]. Available: http://optiwave.com.
30. C. Li, T. G. Ning, C. Zhang, J. Li, C. B. Zhang, X. D. Wen, H. Lin, and L. Pei, “All-fiber multipath Mach-Zehnder interferometer based on a four-core fiber for sensing applications,” Sens. Actuators A Phys. 248, 148–154 (2016). [CrossRef]
