1. Introduction

Salinity is the most fundamental parameter for process control in manufacturing industries, protection of ecosystems, and prevention of global warming [1]. High-quality salinity sensing is very important in these applications, which also have extensive applications in biological, food and chemical industry. In particular, fiber-optic salinity sensors have been widely investigated because they have many distinctive advantages over traditional salinity sensors, such as high sensitivity, miniature size, immunity to electromagnetic interference, and ease of multiplexing [2–9]. In general, the refractive index (RI) of saline water changes as a function of its salinity. Thus, most fiber-optic salinity sensors typically measure the RI of the saline water to determine its salinity and therefore a highly sensitive RI fiber sensor should readily be a good candidate for salinity sensing, such as surface Plasmon resonance effect [2, 3], and RI detection with long-period fiber gratings [4, 5], etched FBG [6], two-core fiber Mach-Zehnder interferometer [7], tapered optical fiber [8], and micro-machined Fabry-Perot (FP) interferometer [9]. Although these methods can provide high sensitivity measurements, the sensor fabrication procedures are complicated and the salinity must be converted from RI, which increased the difficulty of the demodulation.

Meanwhile, an increasing number of functional materials coated fiber salinity sensors are proposed, which can directly determine the salinity; however, most of these salinity sensors developed either had very low salinity sensitivity or complicated fabrication procedures [10–13]. Consequently, how to improve the salinity sensitivity became the current main challenge. Recently there has been a report on the polyimide-coated Hi-Bi photonic crystal fiber Sagnac interferometer as salinity sensor with very good sensitivity [13], although this device provides high salinity sensitivity, it is necessary to use special fiber and complicated configuration, and complicated fabrication procedures had been adopted.

In this paper, a novel fiber salinity sensor based on FP interference is experimentally demonstrated. The proposed sensor is composed of a single mode fiber (SMF) splicing a hollow core fiber (HCF). The end-face of HCF is sealed with a section of the polymer and kept an air-gap remained between the SMF and the polyimide (PI) diaphragm. Once the PI curing, the solid edge of the PI diaphragm can fix the air-gap structure and make the width of the air-gap permanent. The fiber end-face and PI diaphragm act as the mirrors of the FP interferometric cavity. Two beams reflected by the two mirrors generate a well-defined interference spectrum. The swelling property of the PI is related to the salinity. As water salinity increases, the PI diaphragm constriction will induce the increase of the width of the air-gap, which causes the red shift of the fringe. The salinity can be detected and evaluated by measuring the wavelength shift. The sensitivity can be enhanced by reducing air cavity length. Also appropriate air cavity length and PI diaphragm length can be selected to solve the cross-sensitivity between temperature and salinity. Distinct from the previous salinity sensors, the proposed device can be used as an ultra-compact and simple salinity sensor with highly sensitivity, which makes it an alternative candidate as a smart sensor in chemical and biological applications.

2. Sensor design and operating principle

Figure 1 shows a schematic design of the proposed PI diaphragm based air-gap FPI salinity sensor. As illustrated in Figs. 1(a) and 1(b), the sensor is constructed by splicing a piece of HCF to a standard SMF and subsequently cleaving the other end of the HCF to a desired length (HCF was cleaved under a microscope). And as shown as Fig. 1(c), the PI coated on the glass substrate with about 20μm thickness, then, the end-face of HCF is filled with a section of the polymer by dipping it in the PI coat and a liquid PI diaphragm formed on the end of the HCF, which kept an air-gap remained between the SMF and the PI diaphragm. After that, we perform the PI curing process. Finally, the spliced face and the solid PI diaphragm form the FP cavity, as shown in Fig. 1(d) and the inset of Fig. 1. Figure 1(e) shows the operating mechanism. Due to the spliced face, light that propagating in the SMF is partially reflected (I₁) by the fiber-air interface at the splicing point because of the Fresnel reflection and partially illuminate into the air hole of the HCF. When the light reaches the PI diaphragm, a fraction of this light (I₂) is reflected back and coupled into the SMF core again at the splicing point where interference occurs among light I₂ and light I₁. The interference mechanism in the proposed FPI is based on the air-gap FP cavity and interference fringes can be measured experimentally by an optics sensing interrogator, and used to sense the salinity.

 

Fig. 1 Proposed salinity sensor’s production sequence and operating mechanism. Inset: Micrograph of the sensor head with 200μm air-gap and 20μm PI diaphragm.

Download Full Size | PDF

The interferometric spectrum of the FPI can be expressed as:

Thus, the wavelength of the dip λ_m is given by:

The HCF used here was Polymicro Technologies^TM fused silica capillary tubing fabricated by Molex, Inc. It is with an outer diameter of 150μm and inner diameter of 75μm. Figure 2(a) shows the reflection spectrum of the FPI with 200μm air-gap and 20μm PI diaphragm, and the free space range δλ is 6nm. The relationship between the free space range δλ and the air-gap interferometer length L is shown in Fig. 2(b), which meets:

 

Fig. 2 (a) Typical interference pattern with 200μm air-gap and 20μm PI diaphragm. (b) Relationship between the free space range and the air-gap interferometer length.

Download Full Size | PDF

The interference spectrum is dominated by the air-gap cavity since the measured fringe spacing depends only on the air-gap. The reflection by the outward end-face of the PI is negligible.

3. Experiments and discussions

3.1. Salinity responses

Due to the swelling property of the PI correlation with salinity, the proposed PI diaphragm based air-gap FPI is employed to measure the salinity, with the purpose of evaluating the effectiveness of the sensor. In this setup, the reflected light is recorded by an optics sensing interrogator with a resolution of 4 pm. Some sensor prototypes with different air-gap cavity length were fabricated for experimental demonstration of the salinity sensing. For salinity measurement, the sensing tip were immersed into NaCl solutions with salinity ranging from 0mol/L to 5.47mol/L, which are prepared by adding different qualities of NaCl into the same quantity of water. Figure 3 shows the reflection spectra of sensors with different air-gap cavity length, when the sensors were immersed into the solution with different salinities. The fringes shifted towards the long wavelength while the salinity increases.

 

Fig. 3 Red shift of the fringes as the salinity increases.

Download Full Size | PDF

Figure 4(a) presents the dip wavelength shifts as a function of the salinity. The dip wavelength increased linearly as the salinity increased. All wavelengths of these dips shift to long wavelengths when the salinity from 0mol/L to 5.47mol/L. As shown in Fig. 4(a), when salinity increases from 0 to 5.47mol/L, a total of 2.5nm, 2.2nm and 0.59nm wavelength shift were observed, and through a linear fitting, the salinity sensitivities of the FPIs were found to be 0.45, 0.39 and 0.11nm/(mol/L), respectively. These sensitivities are higher than the previous reports, which were based on a hydrogel-coated FBG [10, 11] and a PI-coated FBG [12]. The FP interference fiber salinity sensor has good repeatability and reasonable sensitivity which demonstrated the feasibility of this design. One advantage of our proposed salinity sensor is its robust and ultra-compact structure, and so can be used as the miniature sensor. And the ease of fabrication and lack of need for alignment are favorable characteristics of a fiber interferometer in sensing applications. Additionally, in Fig. 3, the fringe contrast is decreased while immersed the sensor head in surrounding liquids; the change is mainly induced by the PI diaphragm deformation caused by the water swelling. We also investigated the relationship between sensitivity and air-gap cavity length, as illustrated in Fig. 4(b). The sensitivity increases as the air cavity length decreasing, it shows that the sensitivity can be enhanced by reducing air-gap cavity length.

 

Fig. 4 Wavelength shift as a function of water salinity (a) and relationship between sensitivity and air-gap cavity length (b).

Download Full Size | PDF

3.2. Temperature responses

Using a lab-scale temperature measurement system, we experimentally calibrated and tested the temperature responses of the FPI sensors. The FPI probes with different air-gap cavity length and 20μm PI diaphragm are put into a programmable tubular furnace (Y-Feng Electrical Furnace Company, SK2-1-12) for temperature testing, and a thermocouple (KSGD-6.3-16Z) is used for temperature calibration near to the FPIs, both of them are placed in the middle of the furnace chamber. We characterized the sensor performance by tracking a dip of the reflection spectrum. Temperature range was decreased from 80°C to 30°C with a step size of 10°C. Figure 5 presents the reflection spectra of two FPI sensors versus temperature changes, which are based on a 170μm air-gap and 20μm PI diaphragm, and a 200μm air-gap and 20μm PI diaphragm, respectively. When the temperature decreased, the reflection spectra of the sensors shifted toward the short wavelength. In Fig. 5, we noticed that the reflection spectra fringe contrast increased with decreasing temperature, which is caused by the Fresnel reflection change from the PI diaphragm thermal deformation. This phenomenon is as analogous to the fringe contrast change caused by the water swelling of PI diaphragm.

 

Fig. 5 Reflection spectra with decreasing temperature.

Download Full Size | PDF

Figure 6(a) indicates the temperature characteristics of the sensors when the sensors immersed in air. The wavelength of the sensor with 200μm air-gap and 20μm PI diaphragm blue shifts as the temperature decreased, however, the sensor with 170μm air-gap and 20μm PI diaphragm shows temperature-independent. When the temperature decreased from 80 to 30°C, from the linear fit, the temperature sensitivity is 9 pm/°C for the FPI with 200μm air-gap and 20μm PI diaphragm. The sensitivity of this FPI sensor primarily depends on the thermal expansion coefficient of the HCF and the PI. We also investigate the temperature dependence of the sensor with 170μm air-gap and 20μm PI diaphragm, which is illustrated in Fig. 6(a). It confirms that the wavelength stays practically constant over the temperature range from 30 to 80 °C. From a linear fitting, the temperature sensitivity of the FPI with 170μm air-gap and 20μm PI diaphragm is 0.78pm/°C. Notably, to our knowledge, due to the expansion coefficient of the polymer section, the temperature sensitivity of the sensor decreases with the decreasing air-gap cavity length, this means that the polymer section is expanded by the thermal effect as temperature increasing that counteracts the incremental length of the air-gap caused by the thermal expansion of the HCF. The 50°C temperature variation in air caused a total wavelength variation in the amount only about 0.04nm. The temperature dependence of the device was small and contributed less than 2% to the total salinity variation over 50°C temperature range. Therefore, the temperature and salinity cross-sensitivity can be solved by choosing 170μm air cavity length and 20μm PI diaphragm. In addition, we investigated the temperature responses when the sensors immersed in water, as shown in Fig. 6(b). The temperature sensitivity measured in water is slightly higher than that measured in air, which induced by the variation of the swelling property with the temperature. However, The 25°C temperature variation in water caused only about 0.07nm wavelength variation. The temperature dependence of the device was less than 3% to the total salinity variation over 25°C temperature range. Finally, we investigated the salinity sensitivity fluctuation versus temperature, as shown in Fig. 6(c). The salinity sensitivity fluctuation was less than 0.01 nm/(mol/L) over the tested temperature range, which demonstrates the salinity sensitivity is constant during temperature variation. Thus, the temperature effect can be ignored on the salinity response of the sensor prototype.

 

Fig. 6 Relationship between temperature and wavelength shift when the sensors immersed in air in air (a) and water (b). (c) The salinity sensitivity fluctuation versus temperature.

Download Full Size | PDF

4. Conclusions

In summary, we proposed and demonstrated a novel fiber salinity sensor based on FP interference. The fabrication process included splicing between HCF and standard SMF, PI coated on the end-face of HCF and the PI curing. The main principle is based on the swelling property of the PI material. As water salinity increases, the PI diaphragm shrinks, this constriction lead to the increase of the air-gap width, which causes the red shift of the fringe. We fabricated salinity sensor prototypes with different air-gap and 20μm PI diaphragm. The salinity sensing is completed by measuring the wavelength shift of the resonance dips in its reflection spectrum. In the salinity range from 0 to 5.47mol/L, a maximum sensitivity 0.45nm/(mol/L) is achieved on a sensor with 170μm air-gap and 20μm PI diaphragm. Finally, we chose 170μm air-gap and 20μm PI diaphragm combination to solve the temperature and salinity cross-sensitivity. Its robust and ultra-compact structure is the greatest advantage of our proposed miniaturized salinity sensor. Easy fabrication and lack of need for alignment are favorable characteristics of a fiber interferometer in sensing applications. Future works will be focused on salinity sensitivity enhancement by optimizing sensor structure, parameters and materials.