Dual-cavity microstructure fiber optic hydrogen sensor based on evaporated Pt/WO3 film was proposed and experimentally explored in this paper, which provides a novel solution to detect high hydrogen concentration (10-30% H2). Dual-cavity microstructure fabricated by splicer is composed of an inner air-cavity and a collapsed photonic crystal fiber cavity. The proposed sensor has the advantages of miniature structure, stable configuration, low cost. Based on three-beam interference model and verification experiments, the compensation function to the fluctuation of light source and fiber loss is proved from the theoretical simulation and experimental investigation. The sensor has a short response time (1min), good repeatability and reliability. Besides, the change of temperature affects the response value of the hydrogen sensor, but the impact can be neglected in 10-30% H2.
© 2015 Optical Society of America
As a new energy of abundant, high efficiency, non-pollution, hydrogen is presently one of the most effective solutions to solve the energy crisis. Hydrogen gas is widely used in many fields, such as new energy vehicles, aerospace, petro-chemical field. However, it is also rather dangerous due to its high diffusivity, low ignition energy and wide explosion concentration range (between 4% and 75% by volume). Thus, the detection of hydrogen concentration is becoming more and more important. Traditional electrochemical hydrogen sensor is not safe because of the possibility of generating electric sparks in service. Due to their outstanding properties of nature immunity, electromagnetic immunity, and remote-operation capability, fiber optic hydrogen sensor is the research hotpot in recent years.
There are two major types of hydrogen sensitive materials, Palladium (Pd [1,2], Pd/Ag , Pd/Ni ) and tungsten oxide (WO3 [5, 6]). The embrittlement effect of Pd caused by the phase transition after hydrogen cycles,makes the fiber optic hydrogen sensor based on Pd difficult to detect high hydrogen concentration . Doping some other metals in pure palladium can suppress this phase transition and therefore deduces the embrittlement effect to some extent, but the sensitivity and accuracy of these sensitive films will decrease . Therefore, Palladium is difficult to fully meet the requirements of practical applications. WO3 is well-suited to aerobic environment (such as the air) because of antioxidant capacity, and Pt/WO3 has excellent gasochromic response. In our previous work [9, 10], Fiber brag grating (FBG) hydrogen gas sensor based on Pt-loaded WO3 coatings prepared by sol-gel method shows a high sensitivity and fast response especially for relatively low hydrogen concentration (down to 200 ppm) due to the nano-platelet structure of WO3. However, the energy released by exothermic reaction leads to the increase of ambient temperature up to hundreds °C under 4% hydrogen concentration , therefore, the FBG hydrogen sensors have the possibility of exploding in high hydrogen concentration atmosphere.
The intensity reflecting micro-mirror sensor has many unique advantages of convenient fabrication, lower cost, good reliability and flexibility, but susceptible to the fluctuation of light source and fiber loss. To eliminate the noise, two types of compensation technologies have been proposed, including double light channels compensation [12–14], single light channels compensation [15, 16]. Double light channels compensation is difficult to ensure the symmetry of the double channels, and therefore to affect the measurement accuracy of sensor, also the large sensing probe and gas chamber adverse to miniaturization. However, single light channels compensation essentially solves this problem. Moreover, single light channels compensation  has a simpler structure, more easily to achieve miniaturization, which achieves self-referencing and the multiplexing capability.
In this work, a self-compensated dual-cavity microstructure fiber optic hydrogen sensor based on evaporated Pt/WO3 film, which applies to detect high hydrogen concentration, was proposed and experimentally explored. The correlation between hydrogen concentration and response value were investigated and analyzed with Fourier transform demodulation.
The sensor probe has a dual-cavity microstructure, which is composed of an inner air-cavity and a collapsed photonic crystal fiber (PCF) cavity as shown in Figs. 1(a)-1(b). A FITEL S177 splicer, with a manual operation splice mode, was used to fabricate the air cavity. Dual-cavity microstructure is formed by splicing a single-mode fiber and an endless single mode photonic crystal fiber (ESMPCF) to form a closed F-P air-cavity, the uncollapsed section of ESMPCF was cleaved, and the length of collapsed region can be adjusted by cutter . SMF-HOF-SMF  hydrogen sensors have the similar structure with dual-cavity microstructure. However, the two fusion operations and the short length of hollow fiber (The fiber is cleaved by a specially designed in-house tool), makes the sensor difficult to manipulate. By contrast, the easier manufacture method and lower fabrication cost are beneficial to mass production of the dual cavity microstructure fiber optic hydrogen sensor.
Pt/WO3 film, as hydrogen sensitive element is prepared. 350 nm WO3 thin films were deposited by thermal evaporation only on the fiber-tip of the collapsed PCF after ultrasonic cleaning for five minutes. Then 15 nm Pt was sputtered on surface of WO3 film in Ar atmosphere (5 × 10−3 mbar) with BESTECH sputtering system. WO3 and Pt films act as the hydrogen transducer and Catalysts, respectively. The hydrogen sensitive film is also investigated and verified by SEM.
Figure 2 shows the configuration of hydrogen sensing experiment, which consists of a broadband light source (SLED, λ: 1310nm.), an optical spectrum analyzer (OSA, YOKOGAWA, AQ6370B), a chamber, a variable optical attenuator (VOA). Since both the optical source power fluctuation and the mechanical perturbation in the light path affect the light intensity accepted by the optical spectrum analyzer, the light intensity can be adjusted by variable optical attenuator (VOA) instead.
3. Operating principle
As shown in Fig. 3, there are three reflection surfaces in the sensor probe (The reflection from the interface between WO3 and Pd is not taken into account, because the thick of WO3 and Pd are far less than the length of the inner air-cavity and the collapsed PCF cavity.), labeled ‘1’, ‘2’, ‘3′, respectively. The lengths of the air cavity and the collapsed PCF are d1 and d2, respectively. The refractive index of the single mode fiber, air cavity, Collapsed PCF, and evaporated Pt/WO3 film are n0, n1, n2 and ng, respectively. Transmission loss factors at air cavity and collapsed PCF are α1 and α2. According to Fresnel formula, the reflectivity R1 and R2 at surface ‘1’ and ‘2’ are both equal to (n0-n1)2/(n0 + n1)2 = 0.034<<1, while the reflectivity of the surface ‘3′ is less than (n2-ng)2/(n2 + ng)2 = 0.032<<1, owing to the refractive index of the evaporated Pt/WO3 ng ≤ 2.1 in the hydrogen gas. Thus, the total reflected field is approximately equal to the sum of the first-order reflected fields from the three surfaces, is given by
Where, E0 is the input field, r01, r12 and r2g are the reflection coefficient of surfaces 1, 2, 3, respectively. t01, t10 are the transmission coefficient of the surface 1; t12, t21 is the transmission coefficient of the surface 2.
From Eq. (1), we obtain the normalized reflection spectrum can be obtained as follows:
Owing to 4πn1d1>>λ in our experiments, we can loosely think that 
Where, λ0 is the center wavelength of the light source. Analogously,
The Fourier transform spectrum of the reflection spectrum in the positive half axis as follows:
Where b1, b2, b3 and b4 is the impulse at w = 0, w = 4πn1d1/λ02, w = 4πn2d2/λ02 and w = 4π(n1d1 + n2d2)/λ02, respectively, and labeled ‘Peak 0’, ‘Peak 1’, ‘Peak 2’ and ‘Peak 3′, respectively.
Normalized reflection spectrum and its fast Fourier transform spectrum have been theoretically simulated, as shown is Figs. 4(a) and 4(b), respectively. The reflective spectrum experimentally obtained from the sensor probe and its fast Fourier transform spectrum as shown in Figs. 4(c) and 4(d), respectively. It can be found that the half optical path difference (OPD) and relative magnitudes of Peak1, Peak2 and Peak3 in Figs. 4(b) and 4(d) are basically the same, respectively. Therefore, the proposed theory model agrees well with experimental results.
From Eq. (6) it can be concluded that when there exists the fluctuation of light source and fiber loss, the amplitude of Peak1, Peak2 and Peak3 would be fluctuated with the same ratio. Secondly, when the refractive index of the evaporated Pt/WO3 film ng is affected by the hydrogen gas, owing to r2g = (n2-ng)/(n2 + ng), the amplitude of Peak2 and Peak3 would be fluctuated with the same ratio, while the amplitude of Peak1 is not affected. Thus, we can compensate the fluctuation of Peak2 and Peak3 by simply taking ratio with Peak1  as
4. Results and discussion
Surface topology and micro-structure of WO3 film by vacuum thermal evaporation were examined. Figure 5(a) shows the SEM image of Pt/WO3 film after hydrogen cycles, it can be found that the coating surface is rather flat as a whole, and the size distributions of tightly bonded WO3 particles (30-50 nm in diameter) are homogeneous, though micro-cracks are clearly found. Flat surface is favorable to the enhancement of the reflectance of surface 3 and then to the improvement of signal to noise ratio. Enormous micro-cracks are also helpful to generate channels and gaps for hydrogen gas diffusion, which contributes to the improvement of coloring and bleaching process . Figure 5(b) is the XRD pattern of WO3, it can be found that there is an obvious XRD peak near 2θ = 33°, which indicates that there is crystalline phase in the evaporated WO3 film. Compared to amorphous, crystalline state structure has denser surface and more stable structure. Dense surface improves the reflected light intensity of the sensor probe; stable structure is helpful to the long term stability of WO3 film.
The hydrogen sensor was exposed to different concentration hydrogen gas, but the light intensity kept constant. Figure 6(a) shows the interference spectrum experimentally obtained from the sensor probe. The slowly varying fringe pattern is originated from the short air-cavity, while the fine fringes shown in the inset come from the long collapsed PCF cavity. Figure 6(b) is the fast Fourier transform spectrum of the interference spectrum. The hydrogen sensor was exposed to different concentration hydrogen gas, the higher the hydrogen concentration, the smaller the amplitude of Peak2 and Peak3, but the amplitude of Peak1 remains about the same. Theoretically, the higher the H2 concentration, the smaller the refractive index is, and the larger the absorbance index is , so the intensity of interference spectrum and the peak value of Peak2 and Peak3 decreases.
Figure 7 plots the curve fitting of peak value ratio (Sc = Peak3/Peak1) of the sensor under different hydrogen concentrations and the error bar has been displayed. It indicates Sc decreases nonlinearly with the increase of hydrogen concentration and the curve fits well in high range of hydrogen concentration, but not so well in low range. The sensitivity of the sensor is higher in 10%-30% concentration hydrogen than below 5% and above 45% H2. The recovery time of the sensor is 30s. The responding time is about 1min in the range of 10%-30% hydrogen concentration, while it is more than 2 min below 6% H2. Theoretically, coloring process includes two steps: adsorption and reaction of H2 with Pt/WO3 on the surface of the film, diffusion of color centers into interior of WO3 . According to chemical dynamics principle and internal diffusion theory, the latter will be the rate-determining step which depends on the hydrogen partial pressure and therefore, the coloring velocity will increase with the hydrogen concentration . The same principle applies to the bleaching process. The responding time of dual-cavity microstructure fiber optic hydrogen sensors is longer than electrochemical hydrogen sensors. Because the dual-cavity microstructure fiber optic hydrogen sensors works at room temperature, which is far lower than the operating temperature of electrochemical hydrogen sensors. Moreover, Operating temperature can be controlled by laser heating technology  to accelerate the rate of reaction.
The verification experiment were conducted in order to check the compensation to the fluctuation of light source and fiber loss, the response value in 0 and 13.89% concentration of hydrogen were measured with adjusting the light intensity by VOA. Figure 8(a) shows the Fast Fourier transform spectrum in same concentration of hydrogen and different light intensity, Fig. 8(b) shows the compensated amplitude in 0 and 13.89% concentration of hydrogen with different light intensity. Compared with Fig. 5, the amplitude all of Peak1, Peak2 and Peak3 get smaller obviously when reducing the light intensity by VOA only, but Sc keeps mostly unchangeable; when the hydrogen sensor was exposed to different concentration hydrogen gas, the higher the hydrogen concentration, the smaller the amplitude of Peak2 and Peak3, but the amplitude of Peak1 remains about the same. This demonstrates that the system noise of hydrogen sensor is low, and the compensation method is feasible.
In the experiment, the background noise is 0.0075. However, there was the intentional the light intensity fluctuation, the system noise increased to 0.015. However, the sensitivity of hydrogen sensor probe is about 1.449 in 10%-30% H2, the system noise can be accepted; but the sensitivity was close to the system noise in below 6% (about 0.719) or above 45% (about 0.180) H2. Thus, the compensation method is feasible in 10%-30% H2, but does not apply to below 5% or above 45% H2. Owing to the hydrogen sensor exposed to air, the variations of air may produce high frequency noise. Therefore, the system noise is slightly large. The high frequency noise cannot be compensated by this method, but may be solved by seal technology. Moreover, in actual measurement, the spectrometer needs appreciable sweep time to get a set of data, during which the sensor system should be stable, namely the fluctuation of light source and fiber loss faster than scan frequency cannot be properly compensated .
The repeatability and stability of the fiber optic hydrogen sensors with evaporated Pt/WO3 film were investigated. Owing to the short respond time and the long scanning period, only static test was conducted. Figure 9(a) shows the 8 cycle reflective spectrum of the hydrogen sensor, it can be found the reflective spectrum basically coincide with each other in air and in 10% H2, respectively, which indicates that the fiber optic hydrogen sensors has a good repeatability and reliability in spectrum. Figure 9(b) shows the response value of hydrogen cycle, it can be found that the fluctuation of response value is 0.004 in airs, 0.010 in 10% H2, which is consistent with the system noise in the verification experiment. Therefore, such sensor has a good repeatability and reliability.
Figure 10 shows the ambient temperature sensitive properties of the sensing probe in 0°C-80°C. Experimental results show the change of ambient temperature has a small effect on the response value of the hydrogen sensor. In 0°C-80°C, the higher the temperature, the smaller the response value of the hydrogen sensors are, and the temperature sensitivity is 8.265*10−4°C−1. Given that temperature difference between day and night is about 10°C, the fluctuation of the response value is 0.008265, which has the same order of the noise caused by the fluctuation of light source and fiber loss. Therefore, the sensitivity to temperature is smaller than to 10-30% H2, but as much as to below 5% or above 45% H2. Consulted Eq. (6), temperature affects the refractive index ng of the evaporated Pt/WO3 film and dual-cavity microstructure, and this method is difficult to compensate the fluctuation caused by temperature. Indeed, the effect of environmental temperature is indeed limit the sensor's detection precision and measurement range, but the impact can be neglected in 10-30% H2. Besides the ambient temperature, the local heat release produced when evaporated Pt/WO3 film absorbed H2, may affect the performance of the sensor. However, the small amount of Pt/WO3 film and its large surface to volume ratio is helpful to heat conduction, a little heat produced is rapidly spread into the air. Thus, the influence from the heat is small. Moreover, the system error includes the influence of the heat.
At present, several kinds of optical fiber hydrogen sensors have been proposed, such as fiber Bragg grating (FBG) sensor, long-period fiber grating sensor (LPGS). Owing to lots of local heat release produced by the chemical reaction between WO3 and high concentration hydrogen, and the embrittlement effect of Pd in high concentration hydrogen, the structure of FBG sensor  is easily damaged that it can’t detect the high concentration hydrogen. LPGS , which causes coupling between the fundamental core modes and cladding modes, is susceptible to external perturbations. Because the cladding mode encounters the environmental change directly, therefore, transmission spectra changes greatly when there is a little bending of LPG, which makes LPGS impractical. However, the compensation function to the fluctuation of external circumstances makes dual-cavity microstructure fiber optic hydrogen sensors immune to the environmental change. Owing to evaporated Pt/WO3 film, dual-cavity microstructure fiber optic hydrogen sensors is suitable for measuring high concentration hydrogen and have better stability.
Self-compensated dual-cavity microstructure fiber optic hydrogen sensor based on evaporated Pt/WO3 film was proposed and experimentally explored in this paper, which provides a novel solution to detect high hydrogen concentration. Dual-cavity microstructure fabricated by splicer is composed of an inner air-cavity and a collapsed PCF cavity, therefore, the sensor has the advantages of simple and stable structure, low cost, easily to miniaturization. Hydrogen sensitive WO3 film was successfully prepared by vacuum thermal evaporation, as catalyst a 14 nm Pt film was sputtered on 350nm WO3 film to improve its sensitivity to hydrogen response. Based on three-beam interference theoretic model and self-compensation verification experiments (adjusting the light intensity by VOA), the compensation function of dual-cavity microstructure fiber optic hydrogen sensor is proved from the theoretical and experimental. In this work, a simple testing system had been set up and the sensing performance of self-compensated microstructure fiber optic hydrogen sensor film including response value curve under different hydrogen concentrations, repeatability, coloring or bleaching velocity, response threshold value, systematic error and the temperature characteristics is studied. The results indicate response value (Sc) decreases nonlinearly with the increase of hydrogen concentration. The sensitivity of the sensor is higher in 10%-30% H2 than below 6% and above 45% H2. In 10%-30% H2, completely responds time (about 1min) and the recovery time (about 30s) are rather short respectively. Besides, temperature has an effect on response value of the hydrogen sensor, but the impact can be neglected in 10-30% H2. Repeated experiments and SEM images show that Pt/WO3 film is stable without any crack or delamination effect after hydrogen cycles, which demonstrates a good repeatability and reliability of the proposed sensor and hydrogen detection system.
This work is finically supported by the National Natural Science Foundation of China (Project Number: 51402228, 51208398, 61290311), and the Natural Science Foundation of Hubei Province (Project Number: 2014CFB260).
References and links
1. M. A. Butler and D. S. Ginley, “Hydrogen sensing with palladium-coated optical fibers,” J. Appl. Phys. 64(7), 3706–3711 (1988). [CrossRef]
2. M. Wang, M. Yang, J. Cheng, J. Dai, M. Yang, and D. N. Wang, “Femtosecond laser fabricated micro Mach-Zehnder interferometer with Pd film as sensing materials for hydrogen sensing,” Opt. Lett. 37(11), 1940–1942 (2012). [CrossRef] [PubMed]
3. R. J. Westerwaal, J. S. A. Rooijmans, L. Leclercq, D. G. Gheorghe, T. Radeva, L. Mooij, T. Mak, L. Polak, M. Slaman, B. Dam, and T. Rasing, “Nanostructured Pd-Au based fiber optic sensors for probing hydrogen concentrations in gas mixtures,” Int. J. Hydrogen Energy 38(10), 4201 (2013). [CrossRef]
4. J. Dai, M. Yang, X. Yu, K. Cao, and J. Liao, “Greatly etched fiber Bragg grating hydrogen sensor with Pd/Ni composite film as sensing material,” Sens. Actuators B Chem. 174, 253–257 (2012). [CrossRef]
5. V. Wittwer, M. Datz, J. Ell, A. Georg, W. Graf, and G. Walze, “Gasochromic windows,” Sol. Energy Mater. Sol. Cells 84(1-4), 305–314 (2004). [CrossRef]
6. M. H. Yaacob, M. Breedon, and Z. K. Kalantar, “Absorption spectral response of nanotextured WO3 thin films with Pt catalyst towards H2,” Sens. Actuators B Chem. 137(1), 115–120 (2009). [CrossRef]
7. F. Zhou, S. J. Qiu, W. Luo, F. Xu, and Y.-Q. Lu, “An all-fiber reflective hydrogen sensor based on a photonic crystal fiber in-line interferometer,” IEEE Sens. J. 14(4), 1133–1136 (2014). [CrossRef]
8. J. Lee, J. S. Noh, S. H. Lee, B. Song, H. Jung, W. Kim, and W. Lee, “Cracked palladium films on an elastomeric substrate for useas hydrogen sensors,” Int. J. Hydrogen Energy 37(9), 7934–7939 (2012). [CrossRef]
9. M. Yang, G. Wang, J. Dai, Z. Yang, Z. Li, Y. Wang, Y. Zhang, and Z. Zhuang, “Fiber Bragg grating sensors with Pt-loaded WO3 coatings for hydrogen concentration detection down to 200 ppm,” Meas. Sci. Technol. 25(11), 114004 (2014). [CrossRef]
10. M. Yang, Z. Yang, J. Dai, and D. Zhang, “Fiber optic hydrogen sensors with sol-gel WO3 coatings,” Sens. Actuators B Chem. 166-167, 632–636 (2012). [CrossRef]
11. N. Q. Ngo, S. Y. Li, R. T. Zheng, S. C. Tjin, and P. Shum, “Electrically tunable dispersion compensator with fixed center wavelength using fiber Bragg grating,” J. Lightwave Technol. 21(6), 1568–1575 (2003). [CrossRef]
12. Y. Liu, Y. P. Chen, H. Song, and G. Zhang, “Modeling analysis and experimental study on the optical fiber hydrogen sensor based on Pd-Y alloy thin film,” Rev. Sci. Instrum. 83(7), 075001 (2012). [CrossRef] [PubMed]
13. L. Cui, Y. Chen, and G. Zhang, “An optical fiber hydrogen sensor with Pd/Ag film,” Opt. Lett. 5(3), 220–223 (2009). [CrossRef]
14. Y. Zhang, S. L. Qi, and Z. Zhuang, “Fiber optic hydrogen sensor resisting temperature interference,” Proc. SPIE 7753, 775369 (2011). [CrossRef]
15. C. Vazquez, J. Montalvo, D. S. Montero, and J. M. S. Pena, “Self-referencing fiber-optic intensity sensor using ring resonators and fiber Bragg gratings,” IEEE Photonics Technol. Lett. 18(22), 2374–2376 (2006). [CrossRef]
16. Z. Yang, M. Zhang, Y. Liao, Q. Tian, Q. Li, Y. Zhang, and Z. Zhuang, “Extrinsic Fabry-Perot interferometric optical fiber hydrogen detection system,” Appl. Opt. 49(15), 2736–2740 (2010). [CrossRef] [PubMed]
17. K. S. Park, Y. H. Kim, J. B. Eom, S. J. Park, M. S. Park, J. H. Jang, and B. H. Lee, “Compact and multiplexible hydrogen gas sensor assisted by self-referencing technique,” Opt. Express 19(19), 18190–18198 (2011). [CrossRef] [PubMed]
18. G. Zhang, M. Yang, and Y. Dai, “Fabry-Perot fiber tip sensor based on an inner air-cavity for refractive index sensing,” Chin. Opt. Lett. S11202, 12 (2013).
19. Z. Wang, Y. Jiang, W. Ding, and R. Gao, “A white-light interferometry for the measurement of high-finesse fiber optic EFPI sensors,” IEEE Photonics Technol. Lett. 26(21), 2138–2141 (2014). [CrossRef]
20. Z. Li, M. Yang, J. Dai, G. Wang, C. Huang, J. Tang, W. Hu, H. Song, and P. Huang, “Optical fiber hydrogen sensor based on evaporated Pt/WO3 film,” Sens. Actuators B Chem. 1(206), 564–569 (2015). [CrossRef]
21. R. Ghosh, M. B. Baker, and R. Lopez, “Optical properties and aging of gasochromic WO3,” Thin Solid Films 518(8), 2247–2249 (2010). [CrossRef]
22. A. Georg, W. Graf, and R. Neumann, “Mechanism of the gasochromic coloration of porous WO3 film,” Solid State Ion. 127(3-4), 319–328 (2000). [CrossRef]
23. M. Buric, T. Chen, and M. Maklad et al.., “Multiplexable low-temperature fiber Bragg grating hydrogen sensors,” IEEE Photonics Technol. Lett. 21(21), 1594–1596 (2009). [CrossRef]
24. G. M. Ma, J. Jiang, C. R. Li, H. T. Song, Y. T. Luo, and H. B. Wang, “Pd/Ag coated fiber Bragg grating sensor for hydrogen monitoring in power transformers,” Rev. Sci. Instrum. 86(4), 045003 (2015). [CrossRef] [PubMed]
25. Y.-H. Kim, M.-J. Kim, M.-S. Park, J.-H. Jang, B.-H. Lee, and K.-T. Kim, “Hydrogen sensor based on a palladium-coated long-period fiber grating pair,” J. Opt. Soc. Korea 12(4), 221–225 (2008). [CrossRef]