1. Introduction

Liquid level sensing is of great importance in the industrial applications, such as fuel storage systems and biochemical processing. Traditional liquid level sensors are based on the mechanical and electrical techniques. However their applicability is restricted in conductive, flammable and explosive environments. In recent years, fiber grating has attracted considerable interest due to its well-established advantages such as cost-effective, compactness, and ease of multiplexing [1], and fiber grating has been utilized to measure various parameters [2–6], including refractive index, biomaterial, acetylene gas, relative humidity, etc. Moreover, the fiber grating offers overwhelming advantages that it can work in the conductive, flammable, and explosive environment, thus fiber grating offers a promising platform for the liquid level sensing. Many types of fiber grating based liquid level sensors have been proposed and demonstrated [7–13]. One type is based on the fiber Bragg grating (FBG) [7,8]. However, in order to make the FBG sensitive to the external liquid level, the FBG is etched using hydrofluoric acid which makes the device more fragile [7]. The FBG is also fixed on the bending cantilever beam to form the liquid level sensor, inducing the system more complicated [8]. Another type is based on the long-period grating (LPG) [9,10]. LPG based liquid level sensor suffers from wide spectral bandwidth and high cross sensitivity to temperature, which limit its sensing performance [9]. Although the sensing performance is improved by using two identical LPGs to form the interferometric liquid level sensor [10], the fabrication procedure is becoming complicated. The other type is based on the tilted fiber Bragg grating (TFBG) [11–13]. However, the measurement range of these TFBG based sensors is limited to the length of the TFBG itself. Moreover, both LPG and TFBG based liquid level sensors have no reflection sensing spectra, making the sensing system more complicated as compared with the reflective sensors.

In this paper, we propose and demonstrate a simple and compact reflective liquid level sensor based on the modes conversion in the thin-core fiber incorporating one TFBG. As shown in Fig. 1(b), a short piece of ultra-high photon-sensitive thin-core fiber (UHNA) containing a TFBG is spliced to the single-mode fiber (SMF). Compared with other liquid level sensors, the proposed sensor has the simple fabrication process and sensing system, and the measurement range is not limited to the length of TFBG itself. Its sensing principle has been theoretically analyzed and its high sensing performance has been experimentally demonstrated.

 

Fig. 1 Schematic configuration of the liquid level sensing setup (a) and the reflective liquid level sensor (b), and the microscope image of the splicing region between two different fibers (c).

Download Full Size | PDF

2. Operation principle

The operation principle of the proposed reflective liquid level sensor is based on the modes conversion in the thin-core fiber incorporating one TFBG. As shown in Fig. 1(b), the incident light follows two trajectories: one portion of the incident light converts to the cladding modes when it reaches the interface between SMF and UHNA and then these cladding modes convert to the backward propagating core mode by the TFBG (black curve); the other portion continues to propagate as the core mode which is then reflected by the TFBG into the backward propagating cladding modes and core mode, where the reflected cladding modes eventually convert to the core mode at the interface between SMF and UHNA (gray curve). During this conversion, the resulted cladding modes in the thin-core fiber section immersed into the liquid can convert to the radiation modes when their effective refractive index (RI) matches the liquid RI, leading to the decrement of the collected cladding modes power. Thus the liquid level can be measured from the collected cladding modes power.

In order to verify how the optical modes convert in the proposed structure, we use the beam propagation method (BPM) (Rsoft Design Group) to analyze the light propagating along the SMF and UHNA. The core diameters of the SMF and UHNA are 8.2 and 2.5 μm, respectively. The cladding diameters of the SMF and UHNA are 125 μm. Figure 2 shows the stimulated amplitude distribution of the light propagating along the SMF and UHNA under different external surrounding conditions. When the surrounding media is air, the simulation result is shown in Fig. 2(a). It is observed that the incident light beam expands to the high-order cladding modes at the interface between the SMF and UHNA (Z = 1000 μm); the excited cladding modes are then reflected by the boundary of the cladding and surrounding medium to propagate along the fiber. Both the cladding modes and core mode reflected by the TFBG (not shown in the unidirectional BMP model) experience a reverse modes conversion in their return trip, leading to the capture of reflected cladding modes and core mode. Then the UHNA is immersed into the RI solution (RI = 1.451) to study the cladding modes conversion. As shown in Fig. 2(b), when the UHNA section (Z = 6000 to 7000 μm) is immersed into the RI solution, the excited cladding modes convert to radiation modes in this section, leading to the decrement of the cladding modes power as shown in the amplitude distribution in Z = 7000 μm. As shown in Fig. 2(c) (UHNA section Z = 4000 to 7000 μm is immersed into the RI solution) and Fig. 2(d) (UHNA section Z = 2000 to 7000 μm is immersed into the RI solution), when the UHNA section immersed into the RI solution increases, more cladding modes convert to the radiation modes and the amplitude distribution in Z = 7000 μm decreases. It means that the cladding modes power will decrease when the liquid level increases. After reflected by the TFBG, the cladding modes experience the similar liquid-level-dependent modes conversion, leading to the further decrement of the collected cladding modes power. Thus the liquid level can be measured from the collected cladding modes power.

 

Fig. 2 Amplitude distribution of the light propagating along the SMF and UHNA when the surrounding media is air (a) and the different UHNA section is immersed into the RI solution. (b) Z = 6000 to 7000 μm, (c) Z = 4000 to 7000 μm, (d) Z = 2000 to 7000 μm.

Download Full Size | PDF

3. Experimental results

The TFBG is inscribed in the thin-core fiber with the ultra-high photon-sensitivity (Nufern UHNA1, core diameter: 2.5 μm, NA: 0.28). Grating fabrication in the thin-core fiber is feasible due to its high photon-sensitivity. A phase mask based fiber grating fabrication platform using a 244 nm Argon laser is employed to fabricate the TFBG, and the tilt angle is 3°. The length of the fabricated TFBG (L_TFBG) is 10 mm. The UHNA is then spliced to the SMF, and the distance between the TFBG and splicing point (L_D) is 50 mm. The microscope image of the splicing region is shown in Fig. 1(c), it shows that there is no any collapse or distortion in the splicing region, enabling its strong mechanical strength and high repeatability. The other end of the UHNA is angle cleaved to avoid end-face reflection. The proposed reflective liquid level sensor is illuminated by a tunable laser (ANDO AQ 4321D) with the ultra-high stability, while the reflected signal is recorded by an optical spectrum analyzer (ANDO AQ 6317B) through a circulator as shown in Fig. 1(a). Both the cladding modes and core mode are collected as shown in Fig. 3.

 

Fig. 3 Reflection spectrum of the proposed reflective liquid level sensor.

Download Full Size | PDF

To study how the liquid RI influences the modes conversion of the proposed liquid level sensor, the reflective TFBG sensor is immersed into a series of glycerine-water solutions with a fixed liquid level (65 mm). We monitor the collected cladding modes power and spectral changes when the external RI changes. Figure 4(a) shows the collected Bragg resonance wavelength and normalized cladding modes power. When the RI of the solutions increases from 1.333 to 1.42, the high-order cladding modes gradually convert to the radiation modes, and disappear from the shorter wavelength as shown in the inset of Fig. 4(a). The power of the collected cladding modes decreases linearly at a rate of −14.2 dB/R.I.U (R.I.U: refractive index unit); When the RI further increases from 1.42 to 1.467, more low-order cladding modes convert to radiation modes and disappear until only the Bragg resonance is left. The power of the collected cladding modes decreases rapidly at a rate of −60.9 dB/R.I.U. Meanwhile, the peak wavelength of the Bragg resonance keeps constant, making it an ideal temperature reference. We also study the temperature effect on the modes conversion of the proposed liquid level sensor. The reflective TFBG sensor is put in a water bath with a fixed liquid level (65 mm), and the temperature is monitored by a commercial thermometer. Figure 4(b) shows the collected Bragg resonance wavelength and normalized cladding modes power. When the temperature increases from 25 °C to 75 °C, the peak wavelength of the Bragg resonance shifts to the longer wavelength at a rate of 11.2 pm/°C. Meanwhile the Bragg resonance is inherent insensitive to the external RI changes, making it an in situ thermometer. As the external temperature varies, the RI of water also changes at a rate of −1.3 × 10⁻⁴ R.I.U/°C [14]. In order to accurately evaluate the temperature effect, the influence on power response due to the RI variation of water is subtracted by assuming a linear RI response at around 1.333. Its experimental results are shown in Fig. 4(b), the power of the collected cladding modes almost stays constant. It means that the modes conversion of the proposed liquid level sensor is temperature independent, thus we can deduce that the temperature has no influence on the liquid level sensing which is also based on the modes conversion. Since the cladding modes power is highly sensitive to the external RI changes and insensitive to the temperature, while at the same time the Bragg wavelength serves as a good temperature reference, the proposed structure also provides a promising platform for chemical sensing and biosensing.

 

Fig. 4 Response of the reflective liquid level sensor versus the external RI (a) and temperature (b). Insets show the measured reflection spectra.

Download Full Size | PDF

To investigate its liquid level response, the reflective liquid level sensor is fixed inside an acrylic channel to facilitate the level change of the liquid (RI = 1.451). This RI value can ensure that the proposed sensor has the high modes conversion sensitivity as demonstrated above. We monitor the collected cladding modes power to observe the liquid level changes. The experimental results are shown in Fig. 5, the cladding modes gradually convert to the radiation modes and disappear when the liquid level increases. The power of the collected cladding modes decreases at a rate of −0.51 dB/mm. When the liquid level reaches 60 mm, all the cladding modes with the RI matches the external solution vanish. There is no any response even when the liquid level further increases. The experimental results show that the measurement range of the proposed liquid level sensor is the sum of TFBG length (L_TFBG) and the distance between the splicing point and the TFBG (L_D) instead of the TFBG itself. Actually, the RI value of the liquid also influences the measurement range of the proposed liquid level sensor. When the RI increases/decreases, the cladding modes vanish more quickly/slowly, leading the measurement range decrease/increase. For a given RI value, the measurement range depends on how long the cladding modes can propagate along the fiber under the liquid level, and the longest length is determinated in such a way that the collected signal is just above the background noise. Meanwhile, the measurement range can be further increased by increasing the input power.

 

Fig. 5 Response of the reflective sensor versus liquid level. Inset shows the measured reflection spectra.

Download Full Size | PDF

4. Conclusion

In conclusion, we have proposed and demonstrated a new reflective liquid level sensor based on the modes conversion in the thin-core fiber incorporating one TFBG. The simple structural configuration not only eases the fabrication requirements but also improves the compactness and portability of the whole sensing system. Theoretical modeling shows the modes conversion in the proposed structure, and its liquid level sensing ability is demonstrated. The experimental results show that the proposed sensor has a high sensitivity to liquid level and low cross-sensitivity to temperature. And the measurement range is not limited to the length of TFBG itself and can be further increased. Compared with other liquid level sensors, the proposed sensor offers various advantages including low cost, strong mechanical strength, ease of fabrication and large measurement range, and thus is very promising for industrial liquid level measurement applications.