1. Introduction

Hydrogen concentrations monitoring is very crucial in many fields such as aerospace, fuel cells, metal smelting and chemical synthesis. In the past decades, optical fiber hydrogen sensor has been a research issue due to its excellent characteristics such as nature safety, small volume and anti-electromagnetic interference. There are several kinds of optical fiber hydrogen sensors, such as evanescent sensor [1], micro-mirror sensor [2–4], interference sensor [5–7], surface plasmon resonance sensor [8, 9], acoustic resonator sensor [10] and fiber Bragg grating sensor (FBG) [11–22]. Among these sensors, micro-mirror and FBG hydrogen sensor have attracted more interest of researchers due to their simple structure and distributed multiplexing capability, respectively. However, the performance of micro-mirror hydrogen sensor is easily interrupted by fluctuation of optical power of the system. The FBG hydrogen sensor, which is based on Pt-loaded WO₃ coating undergoing an exothermic reaction [12–14] in hydrogen atmosphere, cannot work without oxygen. Some special facilities, such as nuclear waste tank [20], may have little oxygen in it and needs monitoring low hydrogen concentration. FBG hydrogen sensor based on Pd or Pd composite film may be more suitable for hydrogen concentration measurement in such facilities, but its sensitivity is still too low to monitor the lower concentration hydrogen. Moreover, the stability of Pd-based sensor cannot be guaranteed due to the oxidation of the hydrogen sensitive film [11, 21]. WO₃/Pt film has better stability and its gasochromic effect can be maintained without oxygen, which makes it an ideal candidate for hydrogen sensing [23]. In our previous work, the hydrogen performance of WO₃/Pt thin films (towards H₂/N₂ mixture) was investigated with dual-path method. Although the film shows good hydrogen responsibility, the hydrogen sensor is still suffering from the intensity fluctuation of optical system. By using self-referencing technique [24, 25], the measurement errors induced by optical power fluctuation can be greatly compensated, which can significantly improve the anti-interference of sensing system.

In this paper, a more compact and simple optical structure, which consists of high-low reflectivity Bragg gratings and novel hydrogen sensitive film, is employed as hydrogen sensing probe. FBGs with different reflectivities can be easily prepared by controlling the exposure time to excimer laser [26]. Optical power of high reflective FBG (reference FBG) is mostly reflected by the FBG, only a little power of reference signal can reach the fiber tip. Therefore, the peak intensity of high reflective FBG is hardly influenced by the hydrogen sensitive film [22]. However, the peak intensity of low reflective FBG is the joint effect of the hydrogen sensitive film and low reflective FBG. The peak intensity of low reflective FBG, which is sensitive to the reflectivity of hydrogen sensitive film deposited on the tip of single mode fiber, is employed for hydrogen characterization. The peak intensity of high reflective FBG is used as reference. Since Pd composite film has good selectivity towards hydrogen [27, 28] and Pt has excellent stability in air [23], 20 nm Pd₂Pt and 10 nm Pt thin films are prepared as the catalyst layer. WO₃-Pd₂Pt-Pt composite films were deposited on the end-face of singe mode fiber as sensing element, and the sensing characteristics of the hydrogen sensor have been investigated in this work.

2. Experiment

A Lambda Physik excimer laser (COMPex-150T) operating at 248 nm was used as UV light source to write FBG in single mode fiber by phase mask method [29]. FBGs with different reflectivities were prepared by controlling exposure time to excimer laser. Then FBGs was annealed at 100°C for 24 hours to improve their stability. To enhance the stability of FBG, 100 nm Ni thin film was sputtered on the side-face of grating section as a protective layer. The end face of the single mode fiber was firstly deposited with 200 nm WO₃ thin film by thermal evaporation method [23], and oxygen with a flowing rate of 200 standard cubic centimeters per minute (sccm) was supplied as process gas during the evaporation process. After this process, 20 nm Pd₂Pt and 10 nm Pt thin films were sputtered on the surface of WO₃ film by using a BESTECH sputtering system. During the deposition process, the thickness of hydrogen sensitive film was monitored by quartz crystal method. Then single mode fiber inscribed with two Bragg gratings was sealed in porous protective tube by 353ND epoxy glue.

Figure 1 is the schematic diagram of fiber-hydrogen sensor characterization system. The hydrogen sensing performance was carried out at operating temperature of 25°C using nitrogen as carrier gas. Different concentrations of hydrogen were provided by mixing 1% hydrogen (H₂/N₂) and 99.99% nitrogen. The sensing probe was set in a plastic tube with inner diameter of 4 mm, and two kinds of gases were mixed by using two mass flow controls (Beijing Sevenstar, Inc., CS200A, 0~30 sccm and 0~1000 sccm). During hydrogen characterization, the total flowing rate of mixture gas was controlled at 1000 sccm. When hydrogen concentrations were10, 20, 30, 40, 50, 60, 80 and 100 ppm, the flowing rates for H₂/N₂ (1%H₂) were 1, 2, 3, 4, 5, 6, 8 and 10 sccm, and the corresponding flowing rates for N₂ (99.99%) were 999, 998, 997, 996, 995, 994, 992 and 990 sccm respectively. The single mode fiber inscribed with two Bragg gratings was connected to a 10 dBm amplified spontaneous emission light (ASE) and an optical attenuator by a 3 dB coupler. A FBG demodulator module based on InGaAs array detector (Bayspec, Inc. FBGA-F-1525-1565-FA) was used to monitor the central wavelength and peak intensity of FBG. The optical attenuator was employed to ensure the photoelectric detector works at unsaturated state. The collected data is recorded by a computer for further data analysis.

 

Fig. 1 Configuration of optical fiber hydrogen sensor characterization.

Download Full Size | PDF

3. Result and discussion

Figure 2(a) shows the optical spectrum of ASE light source in this work. It can be seen that the ASE has smaller optical intensity variation from 1549 nm to 1554 nm. The principle of this sensor is based on the intensity ratio change of the two FBGs. In order to improve signal noise ratio of the sensing system, the central wavelength of FBG1 and FBG2 should be inscribed in the range of 1549~1554 nm. To obtain the reflective spectrums of the prepared FBGs, the sample was connected to a SLED light (80 μW) and an optical spectrum analyzer (AQ6370B, YOKOGAWA) by a 3 dB coupler. As shown in Fig. 2(b), central wavelengths of high reflective FBG (FBG1) and low reflective FBG (FBG2) are about 1551. 01 nm and 1552.86 nm, respectively. The peak reflectivity of FBG1 and FBG2 are approximately 99% and 15% respectively, and corresponding bandwidths (3 dB) are 0.17 nm and 0.07 nm respectively.

 

Fig. 2 (a) Optical spectrum of ASE light source. (b) Reflective spectrums of single mode fiber inscribed with two FBGs before deposited with hydrogen sensitive film.

Download Full Size | PDF

Figure 3(a) is the reflective spectrum of the sensing probe measured by the FBG demodulator, and obvious interference phenomenon can be seen in the spectrum. The interference fringes, which can be due to Fabry-Perot cavity formed by the FBG2 and the fiber tip, can lead to fluctuations of the peak intensity of the FBGs, resulting in a higher noise level of the sensing single. Thus, the resolution and sensitivity of the hydrogen sensor will be decreased. To overcome this problem, FBG2 with lower reflectivity can be employed to reduce this negative effect. The single noise ratio may be further improved by only inscribing high reflective FBG. As shown in Fig. 3(b), FBG1 and FBG2 have the similar wavelength shifts during 15000 s in air. Figure 3(c) gives the peak intensity change of FBG1 and FBG2 during the noise testing process, and the peak intensity fluctuation of FBG1 and FBG2 is about 0.6%. The variation tendency of peak intensity of two FBGs is not as good as that of their central wavelength, which could be attributed to the optical intensity fluctuation at different wavelength. However, the fluctuation of peak intensity ratio (I₁/I₂) of two FBGs is less than 0.007 (in Fig. 3(d)), which is about 0.3% of I₁/I₂. Therefore, signal noise ratio of the hydrogen sensor can be obviously improved by using I₁/I₂ as measurement parameters. The initial value of I₁/I₂ is around 1.937, which is peak intensity ratio of FBG1 and FBG2 after depositing hydrogen sensitive film on fiber tip.

 

Fig. 3 (a) Reflective spectrum of the sensing probe measured by Bayspec FBG demodulator. (b) Central wavelength and (c) peak intensity of FBG1 and FBG2 for 15000 s at room temperature of 25°C. (d) I₁/I₂ for 15000 s at room temperature of 25°C.

Download Full Size | PDF

Figure 4(a) illustrates the hydrogen response of the sensor under different hydrogen concentrations with nitrogen as carrier gas, and air is used as recovery gas during this process. When hydrogen concentrations are 10, 20, 30, 40, 50, 60, 80 and 100 ppm, the increase of I₁/I₂ are 1.62, 2.25, 2.85, 3.21, 3.5, 3.84, 4.08 and 4.15 respectively, and corresponding response time are approximately 680, 250, 120, 56, 35, 28, 15 and 9 s respectively (as shown in Fig. 4(b)). The increase of the I₁/I₂ value is due to the decrease of I₂, which can be attributed to the decreasing reflectivity of hydrogen sensing film during hydrogen response [22, 23]. The response time decrease significantly with the increase of hydrogen concentrations, which is similar to our previous work [23]. However, the sensitivity of the hydrogen sensor is greatly increased, and the hydrogen sensor has better anti-interference capacity due to its compact structure. The response rate of low concentration hydrogen (below 50 ppm) in nitrogen is much longer, which can be attributed to the slow diffusion of hydrogen molecular under flowing atmosphere. In addition, the relatively lower temperature of the flowing mixture gas, which can reduce heating effect of ASE and self-heating [30] effect during hydrogen response, is responsible for the longer response time of the hydrogen sensor.

 

Fig. 4 (a) Response of the hydrogen sensor under different hydrogen concentrations (H₂/N₂). (b)Response curve under different hydrogen concentrations (H₂/N₂). (c) Increase of I₁/I₂ udder different hydrogen concentrations.

Download Full Size | PDF

As shown in Fig. 4(c), the hydrogen sensor is nearly saturated when the hydrogen concentration reaches to 100 ppm. The error bars represent fluctuation of I₁/I₂ (in Fig. 4(a)) after the hydrogen response reaches equilibrium, and the error bars are not obvious due to the significant increase of I₁/I₂. As the resolution of I₁/I₂ is about 0.001, the hydrogen sensor may detect tens of ppb hydrogen change when the hydrogen concentration ranges from 10 ppm to 60 ppm. In addition, the hydrogen sensor displays ultra-high sensitivity towards low hydrogen in nitrogen atmosphere, which is due to the optimization of hydrogen sensing probe and demodulation technique. Firstly, the better sensitivity of WO₃-Pd₂Pt-Pt composite film in nitrogen results from more generation of WO_3-x [30] or H_xWO₃ [31]. Secondly, the combination of wavelength and intensity demodulation method can ensure a higher signal-noise ratio when central wavelengths of high-low Bragg gratings were specially written in relatively flat range of ASE light source. Compared to the reported self-referencing optical structures [24, 25], the proposed sensing probe can be easily prepared without splicing various optical fibers. Therefore, large optical power loss, which is due to mode field mismatch of different optical fibers, can be avoided by inscribing different reflective Bragg gratings in one single mode fiber.

As it is depicted in Fig. 5(a), the hydrogen sensor exhibits good repeatability during three cycles of 10, 40 and 100 ppm hydrogen mixed with nitrogen, demonstrating good repeatbility of the hydrogen sensing system. After hydrogen response, the morphology of hydrogen sensitive film is shown in Fig. 5(b). The surface of the prepared film looks uniform, and there is no obvious crack on its surface. Therefore, the WO₃-Pd₂Pt-Pt shows good mechanical property. WO₃ has good adhesion towards optical fiber and little volume expansion during hydrogen exposure [13], and lattice constant change of Pd can be greatly suppressed by alloying with Pt [28]. Therefore, the hydrogen sensitive film has good stability during hydrogen exposure.

 

Fig. 5 (a) Repeatbility of the sensor under different hydrogen concentrations (H₂/N₂). (b) SEM of the sensing probe after hydrogen response.

Download Full Size | PDF

4. Conclusion

In summary, a novel optical hydrogen sensor based on fiber-optic self-referenced demodulation technique and WO₃-Pd₂Pt-Pt composite film is proposed and demonstrated. The unique advantage in this method is the employment of both wavelength and intensity demodulation, which enables an ultra-high sensitivity of the proposed hydrogen detection platform. Utilizing the self-referenced demodulation technique and excellent hydro-chromic of WO₃-Pd₂Pt-Pt composite film, we demonstrate detection down to 10 ppm hydrogen and tens of ppb hydrogen change in nitrogen at room temperature. Three key factors (the well-prepared hydrogen sensitive film, the utilization of ASE light source with flat wavelength range and the specially inscribed FBGs) contribute on the excellent performance of the proposed hydrogen sensor. This work discovers new method for developing advanced hydrogen sensing technology with high sensitivity, quick response and good anti-interference, which can greatly broaden its potential application in many hydrogen-related fields. Several methods, such as optimization of hydrogen sensitive films, utilization more stable amplified spontaneous emission light source, lower reflectivity sensing Bragg grating and special wavelength beside high reflective FBG, can be used to improve the performance of the hydrogen sensor. With further improvement, the hydrogen sensor is especially promising for measurement of low concentration hydrogen.