All-glass extrinsic Fabry–Perot interferometer thermo-optic coefficient sensor based on a capillary bridged two fiber ends Zhitao Cao,1 Lan Jiang,1,* Sumei Wang,1 Mengmeng Wang,1,2 Da Liu,1 Peng Wang,1 Fei Zhang,1 and Yongfeng Lu2 1
2
Laser Micro/Nano Fabrication Laboratory, School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China
Department of Electrical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0511, USA *Corresponding author: [email protected] Received 19 January 2015; accepted 9 February 2015; posted 17 February 2015 (Doc. ID 232744); published 18 March 2015
An all-glass extrinsic Fabry–Perot interferometer (EFPI) is demonstrated for thermal-optic coefficient (TOC) of water, glycerol, and their mixture (volume ratio of 1:1). The compensation for the thermal expansion of Fabry–Perot (FP) cavity is realized by assembling a glass capillary and optical fibers through a CO2 laser welding. The thermal responses of EFPIs are tested in air at different cavity lengths of 578.6 μm, 911.7 μm, and 1520.3 μm, respectively. The corresponding refractive index errors induced by thermal expansion of FP cavity are negligible, which are demonstrated to be 4.33 × 10−6 RIU∕°C, 4.13 × 10−6 RIU∕°C, and 3.45 × 10−6 RIU∕°C when temperature increases from 20.03°C to 60.78°C. The thermal-optic coefficients of water, glycerol, and their mixture are measured to be −1.5 × 10−4 RIU∕°C, −2.3 × 10−4 RIU∕°C, and −2.0 × 10−4 RIU∕°C, respectively. Our study suggests a potential use of this sensor for TOC measurements of liquids with the advantages of low costs and robustness. © 2015 Optical Society of America OCIS codes: (120.2230) Fabry-Perot; (060.2370) Fiber optics sensors; (300.6300) Spectroscopy, Fourier transforms. http://dx.doi.org/10.1364/AO.54.002371
1. Introduction
Extrinsic Fabry–Perot interferometric (EFPI) fiber sensors have been widely used to measure temperature, pressure, strain, refractive index (RI), and displacement. In recent years, two-beam interference has been widely used in kinds of FP sensors [1–6]. The simple construction is advantageous to sensor fabrications, and several demodulation algorithms for retrieving the cavity length of an EFPI from the optical spectrum have been proposed. The measurement of RI is rather meaningful to the optical and the chemical investigations of liquids. 1559-128X/15/092371-05$15.00/0 © 2015 Optical Society of America
It can be used to identify substances, or measure the concentrations. RI is strongly dependent on the liquid constitutes and the temperature. Such a RI dependency on temperature is called the thermaloptic coefficient (TOC). A negative TOC of a liquid is a good compensation for thermal expansion of optical devices. A liquid-core is first used to alter the temperature sensitivity of the inscribed gratings in the fiber cladding based on the interaction between the liquid-core with the optical field in cladding [7]. Shuai achieved a reduced thermal sensitivity of 0.67 pm/°C for microfiber Bragg gratings by immersing the microfiber into an ethanol liquid [8]. Xue et al. proposed an isopropanol-sealed optical microfiber taper for high-sensitivity temperature sensing utilizing the high TOC of isopropanol [9]. 20 March 2015 / Vol. 54, No. 9 / APPLIED OPTICS
2371
It is an effective temperature sensitivity enhancement method for extrinsic fiber FP sensor to utilize the high TOC of liquids. An accurate modulation of the TOC determines the temperature compensation effect. Therefore, the measurement of TOC is important for the liquids property investigation and the temperature compensation of optical devices. Several methods have been proposed to explore the TOCs of liquids. Kamikawachi et al. obtained the TOC of ethanol at 1550 nm using etched fiber Bragg gratings [10]. The diameter of an inscribed single mode fiber Bragg grating is etched to be 8 μm, with totally removed fiber cladding for a high RI sensitivity. However, the resonance peak bandwidth is enlarged, resulting in a larger position error of the resonance wavelength. Meanwhile, the robustness is largely weakened because of the corrosion of the fiber cladding, which causes a fiber fragility problem. Viviane et al. reported a method to determining the TOCs of nonfluorescent liquids based on a conical diffraction induced by self-phase modulation [11]. This method required complicated optical alignment adjustments and could be only applied to the nonfluorescent liquids. A two-mode fiber interferometric probe was proposed by Kim et al. to measure the TOC of liquids [12]. The phase and intensity changes of the reflection spectrum are related to the temperature and the surrounding RI variations, respectively. Obviously, the intensity change is largely influenced by the fluctuation of the light source, and an intensity deviation of 0.0052 (a.u.) is measured corresponding to an RI error of 0.001. For FP sensors, the demodulation algorithm is also important for improving the resolution and accuracy of measurement. The demodulation methods such as Fringe visibility monitoring [13], fringe separation detecting [14], and wavelength-tracking [5,15,16] are mainly used to obtain the variation of the measurands. These three methods are all based on the x- or y-coordinate positioning of spectrum fringe peaks or valleys. The automatic identification precision of peak position is very low due to the poor fineness of two-beam interference spectra, reducing the accuracy of interference measurements. Yi Jiang proposed a Fourier-transform white-light interferometry (FTWLI) [17], which was capable of providing both accuracy (0.3∕2300 μm) and stability with a large dynamic measurement range. In this paper, an assembled low-cost fiber optical EFPI sensor was proposed to measure the TOCs of transparent liquids. The FTWLI method was applied to acquire the accurate cavity length of an EFPI for the RI calculation. The accuracy and stability of EFPIs were verified at room temperature. The minimum RI error induced by thermal expansion of FP cavity was 3.45 × 10−6 RIU∕°C, which was two orders of magnitude smaller than the liquid TOCs. Then the cavity length variation within a temperature range from 20.03°C to 60.78°C was investigated in air and three different kinds of transparent liquids. 2372
APPLIED OPTICS / Vol. 54, No. 9 / 20 March 2015
The TOCs of the three kinds of liquids were calculated according to the thermal-based RI change. Finally, a sensitivity-tunable EFPI temperature sensor was constructed using the mixture of glycerol and water. 2. Principle
The schematic diagraph of the EFPI sensor is shown in Fig. 1. The sensor consists of two fiber ends aligned by a silica glass capillary with inner/outer diameters of 0.127/1.8 mm, and one of them is coated with a gold film (about 50 nm) and acts as the second reflector. The gold film was coated on the fiber end by vacuum sputtering. The light from an amplified spontaneous emission (ASE) broadband source was injected into the lead-in optical fiber, and was reflected by the first fiber end and the inner surface of the gold film, respectively. When the medium in the cavity is changed, the effective optical length changes, and this will modulate the FP interference fringes. The interference intensity of I 1 and I 2 against the optical wavelength (λ) can be expressed as p Iλ I 1 λ I 2 λ 2 I 1 λ · I 2 λ cosΔφ1;2 λ; (1) where Δφ1;2 λ 4π λ · L π is the phase difference between I 1 and I 2 , and L is the length of the FP cavity [4]. The FTWLI was applied to recover the cavity length with large dynamic range and high accuracy. The calculated length Lv is the actual cavity length in a vacuum environment. When the determinand is filled into the cavity, the calculated length Lopt is n times longer than Lv , where n is the refractive index of the determinand. Therefore, the RI is measured by
Fig. 1. Schematic and principle diagram of EFPI sensor for TOC detection.
acquiring accurate optical length Lopt of the FP cavity immersed in the determinand. n
Lopt : Lv
(2)
The Lopt is mainly influenced by temperature due to the thermal expansion of FP cavity and the thermal optic change of the determinand. The thermal expansion of the FP cavity is unwanted in the measurement of the TOC, and the thermal optic change is the TOC given by dn∕dT. When values of RI versus different temperatures are measured, the TOC can be calculated. Before assembling with the fibers, the capillary was grinded by a diamond wheel to form a 1000 μm × 1000 μm slot in the middle of the capillary. The 1000 μm depth slot was deep enough to expose the capillary inner tube and used to lead the determinand into the FP cavity for RI measurement. Then the coating-removed fiber end connected to the interrogator was inserted into the capillary. After ideal cutting off and precise assembling, there was a 4% reflection from the cut surface observed from the interrogator. Next, a gold-coated fiber end was inserted into the capillary from the other side. When a proper cavity length was achieved between the two fiber ends, the fibers were welded with the inner sidewall of the capillary using a CO2 laser (SYNRAD firestar V40). The laser power was adjusted to 4 W and focused onto the capillary by a ZnSe lens with a focal length of 50.1 mm. The fibers and the capillary were bonded together at the focal spots after 10 s heating. The distance between the two welding points was 2000 μm. Three sensors with different cavity lengths were fabricated using the same method. The FP cavity length was calculated by subtracting the lengths of two section fibers inserted into the capillary from the length of capillary. The cavity lengths were 578.6, 911.7, and 1520.3 μm, named as EFPI-1, EFPI-2, and EFPI-3, respectively. The glass capillary had a length of 20.2 mm and a thermal expansion of 3.3 × 10−6 ∕°C, which was slightly different from that of the optical fiber as 5.5 × 10−7 ∕°C. The capillary expansion makes a positive contribution to the increase of cavity length, whereas a negative contribution is expected from the fiber expansions. Therefore, the thermal expansion effect can be compensated when the capillary shares the same thermal expansion coefficient with the fibers and both the thermal expansion coefficients are small enough.
Fig. 2. Schematic diagram of the experimental setup for the FTWLI demodulation method.
1 pm, as shown in Fig. 3. The interference spectrum was retrieved for the cavity length calculation with a constant frequency of 2 Hz. Three EFPIs with different cavity lengths were inserted into a glass cylinder, as shown in Fig. 1. The cylinder was heated in a water bath with a temperature accuracy of 0.01°C. The temperature was first maintained at 20.03°C for 166 min to obtain the cavity length fluctuations in air. As shown in Fig. 4, the continuous records of three cavity lengths show that the cavity length fluctuations are less than 0.2 μm. The corresponding RI measurement errors are 3.46 × 10−4 , 2.19 × 10−4 , and 1.32 × 10−4 RIU for EFPI-1, EFPI-2, and EFPI-3, respectively. Actually, the measurement error can be reduced to 10 nm by calculating the average of 40 measurement results at the cost of computational efficiency. For temperature measurements, the EFPIs were heated from 20.03°C to 60.78°C with a step of 10°C. Figure 5 shows the variations of cavity lengths versus temperature for the three EFPIs. The cavity lengths increase by 0.1, 0.15, and 0.21 μm for EFPI-1, EFPI-2, and EFPI-3, corresponding to the RI changes of 4.33 × 10−6 RIU∕°C, 4.13 × 10−6 RIU∕°C, and 3.45 × 10−6 RIU∕°C, respectively. The cavity length increment is mainly induced by the thermal distortion of capillary including axial stretching
3. Experiment
A fiber optical interrogator was utilized to test the EFPI sensor. As shown in Fig. 2, it mainly consists of a broadband amplified spontaneous emission (ASE) source, a fiber Fabry–Perot tunable filter (FFP-TF), two 2 × 2 couplers and two photo detectors (PD). The optical spectrum of the sensor was obtained from 1525 to 1565 nm with a resolution of
Fig. 3. Optical spectrum of EFPI-1 in air and in water. 20 March 2015 / Vol. 54, No. 9 / APPLIED OPTICS
2373
Fig. 4. Cavity length fluctuations of three EFPIs for 166 min. Fig. 6. TOCs of the water, glycerol, and their mixture measured by EFPI-1.
and radial bending. The temperature sensitivity in air can be effectively compensated by reducing the FP cavity length. Because of the smallest cavity length change with temperature, EFPI-1 was chosen to test the temperature caused RI change of water, glycerol, and their mixture from 20.03°C to 60.78°C. The measured liquids were filled into the glass cylinder, and the FP cavity was submerged. In the liquid cavity, the quality of spectrum became worse than in air with the increase of RI, and the cavity length fluctuation of 0.4 μm was obtained at the same temperature as in air. It is noticeable that the expansion of measured liquid does not influence the cavity length because the liquid was not sealed in the cavity. The Lv was replaced by the cavity length in air, because the RI difference of vacuum and air can be ignored. The RIs at different temperature were calculated through dividing the liquid cavity length by the air cavity length at the corresponding temperature. RI decreases were observed with the increases of liquids temperature, as shown in Fig. 6, corresponding to negative TOCs of the liquids. The sensitivities of RI versus temperature, which was the TOC, were
Fig. 5. Cavity length variations versus temperature for three EFPIs. 2374
APPLIED OPTICS / Vol. 54, No. 9 / 20 March 2015
−1.5 × 10−4 RIU∕°C, −2.3 × 10−4 RIU∕°C, and −2.0 × 10−4 RIU∕°C for water, glycerol, and their mixture (volume ratio of 1∶1), respectively. In a previous research, a RI decrease of 0.0060 in the near infrared were proved for water at atmospheric pressure with temperature increasing from 20°C to 60°C [18], corresponding to the TOC of −1.50 × 10−4 RIU∕°C, which is consistent with our results. Most liquids have negative TOCs, and the value is larger than that of optical fibers. Thus, the temperature sensitivity of FP temperature sensors can be enhanced significantly by filling the cavity with high TOC mediums. It is demonstrated that the TOC of glycerol is 1.5 times larger than that of water in our results. However, the RI of glycerol (1.4746) is very close to the RI of fiber core (1.4712), resulting in a poor reflection at the first fiber end, and subsequently reduces the contrast of spectrum fringe. The contrast of spectrum fringe affects the accuracy and stability of the cavity length measurement. Besides, the TOC of the mixture just slightly declines compared with the pure glycerol, whereas the spectrum quality is much better. Hence, the mixture was chosen as sensitivity enhancement medium considering the temperature sensitivity and the accuracy comprehensively. The mixture (1∶1) of glycerol and water was filled into the FP cavity to enhance the temperature sensitivity and the liquid-filled EFPI-1 was heated again from 20.03°C to 70.82°C with a step of 2°C. When it cooled down to 20.03°C, the cavity length was almost identical to the original value, demonstrating a fine sensor repeatability. The detailed temperature measurement result with a good linearity is presented in Fig. 7, and the temperature sensitivity is increased to −0.128 μm∕°C, which is 51 times bigger than that of air-filled EFPI-1. In addition, the temperature sensitivity of the liquid-filled EFPI-1 is adjustable from −0.131 to −0.085 μm∕°C when the ratio of glycerol and water changes from 1:0 to 0:1. We can choose liquids with appropriate TOCs for different demands of temperature sensitivity.
Fig. 7. Cavity length versus temperature for the liquid-filled EFPI-1.
4. Conclusion
In conclusion, the FTWLI method was used to determine the TOCs of transparent liquids using EFPIs. The all-glass structures were simply constructed by a CO2 laser welding of a capillary with two fiber ends to realize the temperature compensation. The maximum cavity length variation in air is 0.2 μm at the temperature of 20.03°C. On the other hand, the TOCs were measured to be −1.5 × 10−4 RIU∕°C, −2.3 × 10−4 RIU∕°C, and −2.0 × 10−4 RIU∕°C with temperature varied from 20.03°C to 60.78°C for water, glycerol, and their mixture, respectively. Moreover, the minimum error induced by thermal expansion of FP cavity was just 3.45 × 10−6 RIU∕°C within the same temperature range. Based on this result, the 1∶1 mixture of glycerol and water is utilized to improve the temperature sensitivity of an EFPI temperature sensor, and the sensitivity enhancement is tunable by changing the ratio of glycerol and water. This research is supported by the National Basic Research Program of China (973 Program) (Grant No. 2011CB013000) and National Natural Science Foundation of China (NSFC) (Grant No. 91323301, 51105038). References 1. L. X. Chen, X. G. Huang, J. Y. Li, and Z. B. Zhong, “Simultaneous measurement of refractive index and temperature by integrating an external Fabry–Perot cavity with a fiber Bragg grating,” Rev. Sci. Instrum. 83, 053113 (2012). 2. 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 Photon. Technol. Lett. 23, 450–452 (2011).
3. J. Tian, Y. Lu, Q. Zhang, and M. Han, “Microfluidic refractive index sensor based on an all-silica in-line Fabry–Perot interferometer fabricated with microstructured fibers,” Opt. Express 21, 6633–6639 (2013). 4. M. Tian, P. Lu, L. Chen, D. Liu, and M. Yang, “Micro multicavity Fabry–Pérot interferometers sensor in SMFs machined by femtosecond laser,” IEEE Photon. Technol. Lett. 25, 1609–1612 (2013). 5. L. Yuan, J. Huang, X. Lan, H. Wang, L. Jiang, and H. Xiao, “All-in-fiber optofluidic sensor fabricated by femtosecond laser assisted chemical etching,” Opt. Lett. 39, 2358–2361 (2014). 6. B. Sun, Y. Wang, J. Qu, C. Liao, G. Yin, J. He, J. Zhou, J. Tang, S. Liu, Z. Li, and Y. Liu, “Simultaneous measurement of pressure and temperature by employing Fabry–Perot interferometer based on pendant polymer droplet,” Opt. Express 23, 1906–1911 (2015). 7. R. Kashyap, D. Williams, and R. P. Smith, “Novel liquid and liquid crystal cored optical fibre Bragg gratings,” in Optical Society of America Topical meeting on Photosensitivity and Quadratic Nonlinearity in Glass Waveguides: Fundamentals and Applications (Optical Society of America, 1997), pp. 25–27. 8. S. Gao, L. Jin, Y. Ran, L. P. Sun, J. Li, and B. O. Guan, “Temperature compensated microfiber Bragg gratings,” Opt. Express 20, 18281–18286 (2012). 9. Y. Xue, Y.-S. Yu, R. Yang, C. Wang, C. Chen, J.-C. Guo, X.-Y. Zhang, C.-C. Zhu, and H.-B. Sun, “Ultrasensitive temperature sensor based on an isopropanol-sealed optical microfiber taper,” Opt. Lett. 38, 1209–1211 (2013). 10. R. C. Kamikawachi, I. Abe, A. S. Paterno, H. J. Kalinowski, M. Muller, J. L. Pinto, and J. L. Fabris, “Determination of thermooptic coefficient in liquids with fiber Bragg grating refractometer,” Opt. Commun. 281, 621–625 (2008). 11. P. Viviane, M. Egberto, and R. R. G. Marcos, “Measurement of the thermo-optic coefficient in liquids by laser-induced conical diffraction and thermal lens techniques,” J. Opt. A 11, 105201 (2009). 12. Y. H. Kim, S. J. Park, S.-W. Jeon, S. Ju, C.-S. Park, W.-T. Han, and B. H. Lee, “Thermo-optic coefficient measurement of liquids based on simultaneous temperature and refractive index sensing capability of a two-mode fiber interferometric probe,” Opt. Express 20, 23744–23754 (2012). 13. M. Jiang, Q.-S. Li, J.-N. Wang, Z. Jin, Q. Sui, Y. Ma, J. Shi, F. Zhang, L. Jia, and W.-G. Yao, “TiO2 nanoparticle thin film-coated optical fiber Fabry–Perot sensor,” Opt. Express 21, 3083–3090 (2013). 14. C. Liao, T. Hu, and D. Wang, “Optical fiber Fabry–Perot interferometer cavity fabricated by femtosecond laser micromachining and fusion splicing for refractive index sensing,” Opt. Express 20, 22813–22818 (2012). 15. D.-W. Duan, Y.-J. Rao, and T. Zhu, “High sensitivity gas refractometer based on all-fiber open-cavity Fabry–Perot interferometer formed by large lateral offset splicing,” J. Opt. Soc. Am. B 29, 912–915 (2012). 16. Y. Wang, D. Wang, C. Liao, T. Hu, J. Guo, and H. Wei, “Temperature-insensitive refractive index sensing by use of micro Fabry–Pérot cavity based on simplified hollow-core photonic crystal fiber,” Opt. Lett. 38, 269–271 (2013). 17. J. Yi, “Fourier transform white-light interferometry for the measurement of fiber-optic extrinsic Fabry–Pérot interferometric sensors,” IEEE Photon. Technol. Lett. 20, 75–77 (2008). 18. A. H. Harvey, J. S. Gallagher, and J. M. H. L. Sengers, “Revised formulation for the refractive index of water and steam as a function of wavelength, temperature and density,” J. Phys. Chem. Ref. Data 27, 761–774 (1998).
20 March 2015 / Vol. 54, No. 9 / APPLIED OPTICS
2375
2
Laser Micro/Nano Fabrication Laboratory, School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China
Department of Electrical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0511, USA *Corresponding author: [email protected] Received 19 January 2015; accepted 9 February 2015; posted 17 February 2015 (Doc. ID 232744); published 18 March 2015
An all-glass extrinsic Fabry–Perot interferometer (EFPI) is demonstrated for thermal-optic coefficient (TOC) of water, glycerol, and their mixture (volume ratio of 1:1). The compensation for the thermal expansion of Fabry–Perot (FP) cavity is realized by assembling a glass capillary and optical fibers through a CO2 laser welding. The thermal responses of EFPIs are tested in air at different cavity lengths of 578.6 μm, 911.7 μm, and 1520.3 μm, respectively. The corresponding refractive index errors induced by thermal expansion of FP cavity are negligible, which are demonstrated to be 4.33 × 10−6 RIU∕°C, 4.13 × 10−6 RIU∕°C, and 3.45 × 10−6 RIU∕°C when temperature increases from 20.03°C to 60.78°C. The thermal-optic coefficients of water, glycerol, and their mixture are measured to be −1.5 × 10−4 RIU∕°C, −2.3 × 10−4 RIU∕°C, and −2.0 × 10−4 RIU∕°C, respectively. Our study suggests a potential use of this sensor for TOC measurements of liquids with the advantages of low costs and robustness. © 2015 Optical Society of America OCIS codes: (120.2230) Fabry-Perot; (060.2370) Fiber optics sensors; (300.6300) Spectroscopy, Fourier transforms. http://dx.doi.org/10.1364/AO.54.002371
1. Introduction
Extrinsic Fabry–Perot interferometric (EFPI) fiber sensors have been widely used to measure temperature, pressure, strain, refractive index (RI), and displacement. In recent years, two-beam interference has been widely used in kinds of FP sensors [1–6]. The simple construction is advantageous to sensor fabrications, and several demodulation algorithms for retrieving the cavity length of an EFPI from the optical spectrum have been proposed. The measurement of RI is rather meaningful to the optical and the chemical investigations of liquids. 1559-128X/15/092371-05$15.00/0 © 2015 Optical Society of America
It can be used to identify substances, or measure the concentrations. RI is strongly dependent on the liquid constitutes and the temperature. Such a RI dependency on temperature is called the thermaloptic coefficient (TOC). A negative TOC of a liquid is a good compensation for thermal expansion of optical devices. A liquid-core is first used to alter the temperature sensitivity of the inscribed gratings in the fiber cladding based on the interaction between the liquid-core with the optical field in cladding [7]. Shuai achieved a reduced thermal sensitivity of 0.67 pm/°C for microfiber Bragg gratings by immersing the microfiber into an ethanol liquid [8]. Xue et al. proposed an isopropanol-sealed optical microfiber taper for high-sensitivity temperature sensing utilizing the high TOC of isopropanol [9]. 20 March 2015 / Vol. 54, No. 9 / APPLIED OPTICS
2371
It is an effective temperature sensitivity enhancement method for extrinsic fiber FP sensor to utilize the high TOC of liquids. An accurate modulation of the TOC determines the temperature compensation effect. Therefore, the measurement of TOC is important for the liquids property investigation and the temperature compensation of optical devices. Several methods have been proposed to explore the TOCs of liquids. Kamikawachi et al. obtained the TOC of ethanol at 1550 nm using etched fiber Bragg gratings [10]. The diameter of an inscribed single mode fiber Bragg grating is etched to be 8 μm, with totally removed fiber cladding for a high RI sensitivity. However, the resonance peak bandwidth is enlarged, resulting in a larger position error of the resonance wavelength. Meanwhile, the robustness is largely weakened because of the corrosion of the fiber cladding, which causes a fiber fragility problem. Viviane et al. reported a method to determining the TOCs of nonfluorescent liquids based on a conical diffraction induced by self-phase modulation [11]. This method required complicated optical alignment adjustments and could be only applied to the nonfluorescent liquids. A two-mode fiber interferometric probe was proposed by Kim et al. to measure the TOC of liquids [12]. The phase and intensity changes of the reflection spectrum are related to the temperature and the surrounding RI variations, respectively. Obviously, the intensity change is largely influenced by the fluctuation of the light source, and an intensity deviation of 0.0052 (a.u.) is measured corresponding to an RI error of 0.001. For FP sensors, the demodulation algorithm is also important for improving the resolution and accuracy of measurement. The demodulation methods such as Fringe visibility monitoring [13], fringe separation detecting [14], and wavelength-tracking [5,15,16] are mainly used to obtain the variation of the measurands. These three methods are all based on the x- or y-coordinate positioning of spectrum fringe peaks or valleys. The automatic identification precision of peak position is very low due to the poor fineness of two-beam interference spectra, reducing the accuracy of interference measurements. Yi Jiang proposed a Fourier-transform white-light interferometry (FTWLI) [17], which was capable of providing both accuracy (0.3∕2300 μm) and stability with a large dynamic measurement range. In this paper, an assembled low-cost fiber optical EFPI sensor was proposed to measure the TOCs of transparent liquids. The FTWLI method was applied to acquire the accurate cavity length of an EFPI for the RI calculation. The accuracy and stability of EFPIs were verified at room temperature. The minimum RI error induced by thermal expansion of FP cavity was 3.45 × 10−6 RIU∕°C, which was two orders of magnitude smaller than the liquid TOCs. Then the cavity length variation within a temperature range from 20.03°C to 60.78°C was investigated in air and three different kinds of transparent liquids. 2372
APPLIED OPTICS / Vol. 54, No. 9 / 20 March 2015
The TOCs of the three kinds of liquids were calculated according to the thermal-based RI change. Finally, a sensitivity-tunable EFPI temperature sensor was constructed using the mixture of glycerol and water. 2. Principle
The schematic diagraph of the EFPI sensor is shown in Fig. 1. The sensor consists of two fiber ends aligned by a silica glass capillary with inner/outer diameters of 0.127/1.8 mm, and one of them is coated with a gold film (about 50 nm) and acts as the second reflector. The gold film was coated on the fiber end by vacuum sputtering. The light from an amplified spontaneous emission (ASE) broadband source was injected into the lead-in optical fiber, and was reflected by the first fiber end and the inner surface of the gold film, respectively. When the medium in the cavity is changed, the effective optical length changes, and this will modulate the FP interference fringes. The interference intensity of I 1 and I 2 against the optical wavelength (λ) can be expressed as p Iλ I 1 λ I 2 λ 2 I 1 λ · I 2 λ cosΔφ1;2 λ; (1) where Δφ1;2 λ 4π λ · L π is the phase difference between I 1 and I 2 , and L is the length of the FP cavity [4]. The FTWLI was applied to recover the cavity length with large dynamic range and high accuracy. The calculated length Lv is the actual cavity length in a vacuum environment. When the determinand is filled into the cavity, the calculated length Lopt is n times longer than Lv , where n is the refractive index of the determinand. Therefore, the RI is measured by
Fig. 1. Schematic and principle diagram of EFPI sensor for TOC detection.
acquiring accurate optical length Lopt of the FP cavity immersed in the determinand. n
Lopt : Lv
(2)
The Lopt is mainly influenced by temperature due to the thermal expansion of FP cavity and the thermal optic change of the determinand. The thermal expansion of the FP cavity is unwanted in the measurement of the TOC, and the thermal optic change is the TOC given by dn∕dT. When values of RI versus different temperatures are measured, the TOC can be calculated. Before assembling with the fibers, the capillary was grinded by a diamond wheel to form a 1000 μm × 1000 μm slot in the middle of the capillary. The 1000 μm depth slot was deep enough to expose the capillary inner tube and used to lead the determinand into the FP cavity for RI measurement. Then the coating-removed fiber end connected to the interrogator was inserted into the capillary. After ideal cutting off and precise assembling, there was a 4% reflection from the cut surface observed from the interrogator. Next, a gold-coated fiber end was inserted into the capillary from the other side. When a proper cavity length was achieved between the two fiber ends, the fibers were welded with the inner sidewall of the capillary using a CO2 laser (SYNRAD firestar V40). The laser power was adjusted to 4 W and focused onto the capillary by a ZnSe lens with a focal length of 50.1 mm. The fibers and the capillary were bonded together at the focal spots after 10 s heating. The distance between the two welding points was 2000 μm. Three sensors with different cavity lengths were fabricated using the same method. The FP cavity length was calculated by subtracting the lengths of two section fibers inserted into the capillary from the length of capillary. The cavity lengths were 578.6, 911.7, and 1520.3 μm, named as EFPI-1, EFPI-2, and EFPI-3, respectively. The glass capillary had a length of 20.2 mm and a thermal expansion of 3.3 × 10−6 ∕°C, which was slightly different from that of the optical fiber as 5.5 × 10−7 ∕°C. The capillary expansion makes a positive contribution to the increase of cavity length, whereas a negative contribution is expected from the fiber expansions. Therefore, the thermal expansion effect can be compensated when the capillary shares the same thermal expansion coefficient with the fibers and both the thermal expansion coefficients are small enough.
Fig. 2. Schematic diagram of the experimental setup for the FTWLI demodulation method.
1 pm, as shown in Fig. 3. The interference spectrum was retrieved for the cavity length calculation with a constant frequency of 2 Hz. Three EFPIs with different cavity lengths were inserted into a glass cylinder, as shown in Fig. 1. The cylinder was heated in a water bath with a temperature accuracy of 0.01°C. The temperature was first maintained at 20.03°C for 166 min to obtain the cavity length fluctuations in air. As shown in Fig. 4, the continuous records of three cavity lengths show that the cavity length fluctuations are less than 0.2 μm. The corresponding RI measurement errors are 3.46 × 10−4 , 2.19 × 10−4 , and 1.32 × 10−4 RIU for EFPI-1, EFPI-2, and EFPI-3, respectively. Actually, the measurement error can be reduced to 10 nm by calculating the average of 40 measurement results at the cost of computational efficiency. For temperature measurements, the EFPIs were heated from 20.03°C to 60.78°C with a step of 10°C. Figure 5 shows the variations of cavity lengths versus temperature for the three EFPIs. The cavity lengths increase by 0.1, 0.15, and 0.21 μm for EFPI-1, EFPI-2, and EFPI-3, corresponding to the RI changes of 4.33 × 10−6 RIU∕°C, 4.13 × 10−6 RIU∕°C, and 3.45 × 10−6 RIU∕°C, respectively. The cavity length increment is mainly induced by the thermal distortion of capillary including axial stretching
3. Experiment
A fiber optical interrogator was utilized to test the EFPI sensor. As shown in Fig. 2, it mainly consists of a broadband amplified spontaneous emission (ASE) source, a fiber Fabry–Perot tunable filter (FFP-TF), two 2 × 2 couplers and two photo detectors (PD). The optical spectrum of the sensor was obtained from 1525 to 1565 nm with a resolution of
Fig. 3. Optical spectrum of EFPI-1 in air and in water. 20 March 2015 / Vol. 54, No. 9 / APPLIED OPTICS
2373
Fig. 4. Cavity length fluctuations of three EFPIs for 166 min. Fig. 6. TOCs of the water, glycerol, and their mixture measured by EFPI-1.
and radial bending. The temperature sensitivity in air can be effectively compensated by reducing the FP cavity length. Because of the smallest cavity length change with temperature, EFPI-1 was chosen to test the temperature caused RI change of water, glycerol, and their mixture from 20.03°C to 60.78°C. The measured liquids were filled into the glass cylinder, and the FP cavity was submerged. In the liquid cavity, the quality of spectrum became worse than in air with the increase of RI, and the cavity length fluctuation of 0.4 μm was obtained at the same temperature as in air. It is noticeable that the expansion of measured liquid does not influence the cavity length because the liquid was not sealed in the cavity. The Lv was replaced by the cavity length in air, because the RI difference of vacuum and air can be ignored. The RIs at different temperature were calculated through dividing the liquid cavity length by the air cavity length at the corresponding temperature. RI decreases were observed with the increases of liquids temperature, as shown in Fig. 6, corresponding to negative TOCs of the liquids. The sensitivities of RI versus temperature, which was the TOC, were
Fig. 5. Cavity length variations versus temperature for three EFPIs. 2374
APPLIED OPTICS / Vol. 54, No. 9 / 20 March 2015
−1.5 × 10−4 RIU∕°C, −2.3 × 10−4 RIU∕°C, and −2.0 × 10−4 RIU∕°C for water, glycerol, and their mixture (volume ratio of 1∶1), respectively. In a previous research, a RI decrease of 0.0060 in the near infrared were proved for water at atmospheric pressure with temperature increasing from 20°C to 60°C [18], corresponding to the TOC of −1.50 × 10−4 RIU∕°C, which is consistent with our results. Most liquids have negative TOCs, and the value is larger than that of optical fibers. Thus, the temperature sensitivity of FP temperature sensors can be enhanced significantly by filling the cavity with high TOC mediums. It is demonstrated that the TOC of glycerol is 1.5 times larger than that of water in our results. However, the RI of glycerol (1.4746) is very close to the RI of fiber core (1.4712), resulting in a poor reflection at the first fiber end, and subsequently reduces the contrast of spectrum fringe. The contrast of spectrum fringe affects the accuracy and stability of the cavity length measurement. Besides, the TOC of the mixture just slightly declines compared with the pure glycerol, whereas the spectrum quality is much better. Hence, the mixture was chosen as sensitivity enhancement medium considering the temperature sensitivity and the accuracy comprehensively. The mixture (1∶1) of glycerol and water was filled into the FP cavity to enhance the temperature sensitivity and the liquid-filled EFPI-1 was heated again from 20.03°C to 70.82°C with a step of 2°C. When it cooled down to 20.03°C, the cavity length was almost identical to the original value, demonstrating a fine sensor repeatability. The detailed temperature measurement result with a good linearity is presented in Fig. 7, and the temperature sensitivity is increased to −0.128 μm∕°C, which is 51 times bigger than that of air-filled EFPI-1. In addition, the temperature sensitivity of the liquid-filled EFPI-1 is adjustable from −0.131 to −0.085 μm∕°C when the ratio of glycerol and water changes from 1:0 to 0:1. We can choose liquids with appropriate TOCs for different demands of temperature sensitivity.
Fig. 7. Cavity length versus temperature for the liquid-filled EFPI-1.
4. Conclusion
In conclusion, the FTWLI method was used to determine the TOCs of transparent liquids using EFPIs. The all-glass structures were simply constructed by a CO2 laser welding of a capillary with two fiber ends to realize the temperature compensation. The maximum cavity length variation in air is 0.2 μm at the temperature of 20.03°C. On the other hand, the TOCs were measured to be −1.5 × 10−4 RIU∕°C, −2.3 × 10−4 RIU∕°C, and −2.0 × 10−4 RIU∕°C with temperature varied from 20.03°C to 60.78°C for water, glycerol, and their mixture, respectively. Moreover, the minimum error induced by thermal expansion of FP cavity was just 3.45 × 10−6 RIU∕°C within the same temperature range. Based on this result, the 1∶1 mixture of glycerol and water is utilized to improve the temperature sensitivity of an EFPI temperature sensor, and the sensitivity enhancement is tunable by changing the ratio of glycerol and water. This research is supported by the National Basic Research Program of China (973 Program) (Grant No. 2011CB013000) and National Natural Science Foundation of China (NSFC) (Grant No. 91323301, 51105038). References 1. L. X. Chen, X. G. Huang, J. Y. Li, and Z. B. Zhong, “Simultaneous measurement of refractive index and temperature by integrating an external Fabry–Perot cavity with a fiber Bragg grating,” Rev. Sci. Instrum. 83, 053113 (2012). 2. 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 Photon. Technol. Lett. 23, 450–452 (2011).
3. J. Tian, Y. Lu, Q. Zhang, and M. Han, “Microfluidic refractive index sensor based on an all-silica in-line Fabry–Perot interferometer fabricated with microstructured fibers,” Opt. Express 21, 6633–6639 (2013). 4. M. Tian, P. Lu, L. Chen, D. Liu, and M. Yang, “Micro multicavity Fabry–Pérot interferometers sensor in SMFs machined by femtosecond laser,” IEEE Photon. Technol. Lett. 25, 1609–1612 (2013). 5. L. Yuan, J. Huang, X. Lan, H. Wang, L. Jiang, and H. Xiao, “All-in-fiber optofluidic sensor fabricated by femtosecond laser assisted chemical etching,” Opt. Lett. 39, 2358–2361 (2014). 6. B. Sun, Y. Wang, J. Qu, C. Liao, G. Yin, J. He, J. Zhou, J. Tang, S. Liu, Z. Li, and Y. Liu, “Simultaneous measurement of pressure and temperature by employing Fabry–Perot interferometer based on pendant polymer droplet,” Opt. Express 23, 1906–1911 (2015). 7. R. Kashyap, D. Williams, and R. P. Smith, “Novel liquid and liquid crystal cored optical fibre Bragg gratings,” in Optical Society of America Topical meeting on Photosensitivity and Quadratic Nonlinearity in Glass Waveguides: Fundamentals and Applications (Optical Society of America, 1997), pp. 25–27. 8. S. Gao, L. Jin, Y. Ran, L. P. Sun, J. Li, and B. O. Guan, “Temperature compensated microfiber Bragg gratings,” Opt. Express 20, 18281–18286 (2012). 9. Y. Xue, Y.-S. Yu, R. Yang, C. Wang, C. Chen, J.-C. Guo, X.-Y. Zhang, C.-C. Zhu, and H.-B. Sun, “Ultrasensitive temperature sensor based on an isopropanol-sealed optical microfiber taper,” Opt. Lett. 38, 1209–1211 (2013). 10. R. C. Kamikawachi, I. Abe, A. S. Paterno, H. J. Kalinowski, M. Muller, J. L. Pinto, and J. L. Fabris, “Determination of thermooptic coefficient in liquids with fiber Bragg grating refractometer,” Opt. Commun. 281, 621–625 (2008). 11. P. Viviane, M. Egberto, and R. R. G. Marcos, “Measurement of the thermo-optic coefficient in liquids by laser-induced conical diffraction and thermal lens techniques,” J. Opt. A 11, 105201 (2009). 12. Y. H. Kim, S. J. Park, S.-W. Jeon, S. Ju, C.-S. Park, W.-T. Han, and B. H. Lee, “Thermo-optic coefficient measurement of liquids based on simultaneous temperature and refractive index sensing capability of a two-mode fiber interferometric probe,” Opt. Express 20, 23744–23754 (2012). 13. M. Jiang, Q.-S. Li, J.-N. Wang, Z. Jin, Q. Sui, Y. Ma, J. Shi, F. Zhang, L. Jia, and W.-G. Yao, “TiO2 nanoparticle thin film-coated optical fiber Fabry–Perot sensor,” Opt. Express 21, 3083–3090 (2013). 14. C. Liao, T. Hu, and D. Wang, “Optical fiber Fabry–Perot interferometer cavity fabricated by femtosecond laser micromachining and fusion splicing for refractive index sensing,” Opt. Express 20, 22813–22818 (2012). 15. D.-W. Duan, Y.-J. Rao, and T. Zhu, “High sensitivity gas refractometer based on all-fiber open-cavity Fabry–Perot interferometer formed by large lateral offset splicing,” J. Opt. Soc. Am. B 29, 912–915 (2012). 16. Y. Wang, D. Wang, C. Liao, T. Hu, J. Guo, and H. Wei, “Temperature-insensitive refractive index sensing by use of micro Fabry–Pérot cavity based on simplified hollow-core photonic crystal fiber,” Opt. Lett. 38, 269–271 (2013). 17. J. Yi, “Fourier transform white-light interferometry for the measurement of fiber-optic extrinsic Fabry–Pérot interferometric sensors,” IEEE Photon. Technol. Lett. 20, 75–77 (2008). 18. A. H. Harvey, J. S. Gallagher, and J. M. H. L. Sengers, “Revised formulation for the refractive index of water and steam as a function of wavelength, temperature and density,” J. Phys. Chem. Ref. Data 27, 761–774 (1998).
20 March 2015 / Vol. 54, No. 9 / APPLIED OPTICS
2375
