Open Access
Add to BibSonomyAbstract
We report a photonic crystal fiber (PCF) in-line Mach-Zehnder interferometer used as a gas sensor device which exhibits high sensitivity to the explosive trinitrotoluene (TNT). The interferometric sensor head is formed by embedding a segment of large-mode-area/grapefruit PCF between standard single-mode fibers via butt coupling, which produces two small air gaps in between terminated fiber ends with ceramic ferrule connectors as coupling regions, which also serve as inlet/outlet for the gas. The spectral response of the interferometer is investigated in terms of its wavelength spectrum. The selectivity to TNT vapor is achieved by immobilizing a molecular recognition ployallylamine layer on the inner surface of the holey region of the PCF. The TNT-induced variations of the interference fringes are measured and the sensing capability of the proposed sensor is demonstrated experimentally.
© 2016 Optical Society of America
1. Introduction
The detection of explosives and their residues is of great importance in public health, antiterrorism and homeland security applications [1–3]. The vapor pressures of most explosive compounds are extremely low and attenuation of the available vapor is often great due to diffusion in the environment, making direct vapor detection difficult. In reality bomb dogs are still the most efficient way to quickly detect explosives on the spot [4]. Interestingly, detection of TNT (2,4,6-trinitrotoluene) vapor using a waveguide interferometry was proven to be highly selective and sensitive when sol–gel films prepared from alkoxysilane precursors bearing a bridging aromatic group are used [5].
In recent years, many compact and robust in-fiber Mach-Zehnder interferometric sensors based on modal interference in photonic crystal fiber (PCF) have been developed for sensing refractive index, strain, pressure, temperature, and humidity, etc. Such interferometers can be built via several techniques such as tapering, splicing, long period grating inscription, and partially micro-holes collapsing [6–14]. Generally, the low temperature sensitivity is a great merit in photonic crystal fiber modal interferometric sensors [15–17]. However, in order to meet specific requirements in terms of gas inlet/outlet and faster gas flux, particular approaches to build compact modal interferometers with PCFs need to be further explored to make them more attractive for gas sensing applications [18].
In this paper, a photonic-microfluidic integrated sensor for highly sensitive TNT vapor detection in a few parts-per-billion range (ppbv) is described based on an in-line Mach-Zehnder interferometer (MZI) in grapefruit PCF or endlessly single mode, large-mode-area (LMA) PCF. A segment of PCF is inserted between standard single-mode fibers (SMF) via butt coupling to form a modal interferometer, in which the cladding modes are excited and interfere with the fundamental core mode. Due to butt coupling, the small air gap between SMF and PCF forms a coupling region and also serves as an inlet/outlet for the gas. The sensor is fabricated by immobilizing a molecular recognition polymer layer on the inner surface of the holey region of the PCF, which can selectively and reversibly bind TNT molecules on the sensitized surface. The sensing mechanism is based on the determination of the TNT-induced wavelength shift of interference fringes due to the refractive index change of the holey-layer. The sensor device therefore is capable of field operation.
2. Interferometer configuration and spectral characteristics
By employing LMA PCF [Fig. 1(a)] and grapefruit PCF [Fig. 1(b)] respectively, we designed simple and compact PCF in-line Mach-Zehnder interferometers that can be used as gas sensors for explosive detection. Both the LMA PCF and the grapefruit PCF are commercial products. The LMA PCF (Yangtze Optical Fibre and Cable Co., Ltd) designed to endless single mode comprises of 5 layers of air holes around a solid core of a diameter 12 μm. And the pitch is Λ = 8.2 μm with normalized air-hole diameter being d/Λ = 0.52. The core and cladding material is pure silica with a refractive index of 1.444. The grapefruit PCF (FiberHome Technologies Group) has an outer diameter of 125 μm, a core diameter of 6 μm, and six large cladding air holes with a diameter of 30 μm in the radial direction. The refractive indices of germanium-doped core and the bulk silica material surrounding the core are 1.479 and 1.457 for λ = 1550 nm, respectively. The region between the cladding air holes and the core is referred to as the inner cladding with diameter of 16 μm. The schematic of interferometer configuration is shown in Fig. 1(c). To excite the core mode to couple out with cladding modes, two butt couplings were devised to form free space beams. The fiber ends were slid into ceramic ferrule pairs connected with ferrule mating split sleeves. Ferrule mating split sleeves properly align the cores of each terminated fiber end with ceramic ferrule connector and minimize back reflections by using 8° pre-angled ceramic ferrules on the each end of SMFs. The first air gap for SMF to PCF coupling can couple a part of the core mode to certain cladding modes. After the two light beams propagate at different speeds over a certain PCF length of the core and cladding, the cladding modes will be recoupled back to the core when they meet the second air gap for PCF to SMF coupling, thus creating interference between the core and cladding modes. Significantly, the air gaps at both ends of PCF can form the inlet/outlet routes for the gas.
Fig. 1 Cross-sections of (a) LMA PCF and (b) grapefruit PCF. (c) Schematic of PCF–MZI built via butt coupling.
The transmission spectra of the fabricated PCF–MZI device shown in Fig. 2(a) were measured using an AQ6317b Optical Spectrum Analyzer (OSA) together with a broadband source (ASLD-CWDM-5-B-FA, Amonics). In order to see the interference pattern more clearly we need to subtract the fixed light source spectrum (marked as trace A: FIX) from the measured interferometric spectrum (marked as trace B: WRITE). Figures 2(b) and 2(c) show experimental results based on 60 mm-long LMA PCF–MZI and 60 mm-long grapefruit PCF–MZI. High visibility interference fringes were observed over the range of wavelengths. At a wavelength of 1550 nm, the measured loss of the LMA type device was around 10 dB and the optical loss of each butt coupled joint was about 5 dB. Whereas, for the grapefruit type device, the measured loss was around 12 dB and the optical loss of each butt coupled joint was about 6 dB, which was slightly larger than that of the LMA type. In the proposed interferometer, the optical path length of the free-space guiding between SMF and PCF is an important consideration that should be made. We compared the average amplitude of transmission spectra of the interferometer with different working distances. As is shown in Figs. 2(d) and 2(e), we found that a moderate working distance (~200 μm for LMA type and ~250 μm for grapefruit type, respectively) induced larger amplitude, which might be attributed to considerable diffraction and diameter expansion of the free space beam. Therefore, we chose 200 μm and 250 μm long of working distances in the experiments for both types of interferometers, respectively.
Fig. 2 (a) Image of the PCF–MZI device. Measurement of transmission spectra of (b) 60mm-long LMA PCF–MZI and (c) 60mm-long grapefruit PCF–MZI. (d) Gap spacing between terminated fiber ends with ceramic ferrule pairs. (e) The average amplitude of transmission spectra around 1550 nm as a function of the working distance between SMF and PCF.
In order to investigate the dependence of the interference fringe on the physical length of PCF (L), both interferometers with different PCF lengths were fabricated, and normalized transmission spectra by subtracting the DC component are summarized in Figs. 3(a) and 3(b). As the PCF length increases, the fringe spacing of both types of interferences became finer. However, the visibility of interference fringes would decrease as the PCF length L increases. The testing PCF has only a maximum of 180 mm in length during measurement. By comparison, the LMA type tends to form a more uniform interference fringe throughout the entire transmission spectrum. It is quite obvious that, however, the interference fringes of the grapefruit type are constitutive of several spatial frequency components arising from interference between the fundamental core mode and multiple cladding modes. It is difficult to identify the spatial frequency components by looking at the original transmission spectrum. Thus, the normalized amplitudes of interference modes were converted to the spatial frequency domain by taking the fast Fourier transform [19,20], shown in Figs. 3(c) and 3(d). The LMA type has only one dominant peak or two seriously overlapped strong peaks in the spatial frequency spectra for all PCF lengths, while one dominant peak and a few separate and moderate-intensity sub-peaks arise together in the grapefruit type. In principle, different spatial frequency components have to do mainly with the excitation of two or more cladding modes, which might cause the coupling between the fundamental core mode (LPcore) and multiple cladding modes (LPclad,m) of the PCF [19]. The phase difference between the LPcore and LPclad,m modes accumulated during the PCF length can be expressed as , with the difference of effective mode indices . As shown in Fig. 4, the spatial frequency corresponding to the dominant peak increases linearly with the PCF length. The fitted curve for each type of interferometer has a different slope, 0.11 mm−1/cm for LMA type and 0.029 mm−1/cm for grapefruit type, respectively.
Fig. 3 Normalized transmission spectra (arbitrary unit) of (a) the LMA PCF–MZI and (b) the grapefruit PCF–MZI with different PCF lengths. (c),(d) corresponding power spectra in the spatial frequency domain.
Fig. 4 Spatial frequency as a function of the physical length of LMA PCF and grapefruit PCF, respectively. Linear fitting equations with their correlation coefficients were also shown in the figure.
As mentioned above, the recorded interference signal is a function of wavelength rather than time. For the Mach-Zehnder interference between the LPcore mode and the dominant LPclad,m mode of the PCF, the transmission spectrum can be expressed as
where and denote the light field intensities of the LPcore mode and the LPclad,m mode, respectively. One-dimensional Taylor series expansion of the function about a point (a chosen central wavelength) is given byHere, the wavelength deviation . Note that the differential effective modal group index is referred to take into account the waveguide dispersion and the material dispersion of the waveguide. Substitute this result into Eq. (2) to giveThe spatial frequency, denoted by ν, is a measure of how often sinusoidal components (as determined by the Fourier transform) of the interference fringes repeat per unit of wavelength. Compare interference waveform function with Eq. (3) to giveThe spatial frequency ν is linearly proportional to the PCF physical length L. The differential effective mode index can be determined by (refer to the slope of fitting curve in Fig. 4), which could be calculated as 0.0264 RIU for LMA PCF and 0.0070 RIU for grapefruit PCF, respectively. Compared with grapefruit PCF, the calculated differential effective mode index for LMA PCF is much higher probably because the high-order cladding mode with a smaller effective index was involved in the interference [20].Based on the discussion above, the selective sensing of specific gases will be realized if a particular gas molecular sensitive thin-film can be coated on the inner surface of the cladding air channels of the PCF in an interferometer. The interferometer is acutely sensitive to the change in refractive index of the interior coating. The effective index of the cladding mode will show a tiny variation depending on the capturing of the gas molecule, which leads to a shift in the interference fringes. The wavelength of the zth order interference valley/dip can be derived as
Due to a combination of factors, the magnitude of the interferometric response to gaseous analyte is far greater than what can be accounted for by pure mass addition [5].3. Explosive sensing applications
3.1 Sensors fabrication and sensing mechanism
In this section, we experimentally present a simple and effective interferometric method for the detection of TNT, which is based on TNT recognition by ployallylamine (PAH) deposited on the inner surface of air holes in the LMA or grapefruit PCF–MZI. This process is shown in Fig. 5. Firstly, a 0.5 mg mL−1 solution of PAH in ethanol was filled into air channels of the PCF by using a syringe. Before that, one end of PCF was inserted into the syringe needle and sealed with a heat shrink sleeve in the heater prior to pulling the plunger back to fill the syringe. The coating solution was left inside for half an hour. Then, pull the plunger back on the syringe to suck out the remaining fluid in the PCF and let the inner surface dry thoroughly at 60°C. This approach led to a PAH coating with a refractive index of ~1.5 and a thickness of 0.5~0.6 μm on the inner surface of holey region in the PCF. SEM photographs in cross section through PCFs are presented in Fig. 6(a). It is noted that the images of PCFs are not very clear due to the poor conductivity of the fibers. The PCF–MZI sensor head was fabricated by embedding a 60 mm-long PCF with PAH internal coating between two SMFs via butt coupling, shown in Figs. 1 and 2.
Fig. 5 Sketch for sensor fabrication and sensing mechanism. (a) The cleaned inner surface of holey region of PCF. (b) Ployallylamine film deposition on the inner surface of holey region of the PCF. (c) Interaction between polyallylamine and TNT molecule to form the Meisenheimer complex.
Fig. 6 (a) SEM images of the cross-section of the fiber coated with polyallylamine film (scale bar 2 μm). (b) Schematic diagram of experimental setup for investigation sensing capability of the proposed interferometric sensor.
In principle, the as-prepared PAH film can be directly used as a recognition probe for the detection of TNT. It is mainly because of the strong interaction between the TNT molecule and the ployallylamine bearing a huge number of amine groups [21]. The resulting formation of Meisenheimer complexes (i.e., TNT–amine complexes [22,23]) could induce variations of the interference fringes by changing the effective index of the cladding mode of PCF. Thus, the trace vaporous TNT can be detected by monitoring the wavelength shifts.
In experiments, a small amount of commercially available TNT was sealed in a 24 mL syringe and heated at 85°C for several hours and then cooled to room temperature to generate volumes of TNT-saturated air. This vaporous TNT in the syringe has a vapor pressure of ~9.15 × 10−9 atm (corresponding to 9.15 ppbv) at 25°C according to [24]. In order to obtain lower testing concentrations the TNT saturated vapor was quantitatively diluted with dry air. Prior to injection, the needle was slid into the air inlet of the PCF–MZI device that clamped at a stage and then sealed with epoxy resin to ensure excellent gas tightness. Continuous vaporous TNT/air flow was injected at the air inlet of the PCF–MZI sensor. The transmission spectra of the sensor were recorded and analyzed in a personal computer as the sensor was exposed to TNT at room temperature. Figure 6(b) shows the schematic diagram of the measurement system. The wavelength resolution of the OSA was set to be 0.01 nm in measurements, unless otherwise specified.
3.2 LMA PCF–MZI sensor response to TNT vapor
Tests indicated that the interferometer device without PAH coating in the PCF was not TNT-sensitive. Figure 7(a) shows the normalized transmission spectra of the interferometric sensor with the PAH-coated LMA PCF measured with dry air and with TNT saturated air. The calculated Q-factor of the LMA PCF–MZI sensor can reach up to 1.3 × 103. Note from the figure that the interference fringe was shifted toward the shorter wavelength direction with the exposure to TNT vapor and a cumulative shift of 1.25 nm is reached. To measure the wavelength variation to TNT concentration, an interference dip centered at ~1550 nm was chosen from the transmission spectrum. The concentration of TNT vapor was increased from 0 to 9.15 ppbv. The wavelength shift is shown in Fig. 7(b). Wavelength blue shift versus TNT concentration is nearly linear, though data points from measurements is sparse. A sensitivity of 140 pm/ppbv can be obtained by using the linear regression fit and a limit of detection of 0.2 ppbv can be draw from the calculated sensor resolution of 27 pm according to three standard deviations (3σ) method [25].
Fig. 7 (a) Normalized transmission spectra (arbitrary unit) of the LMA PCF–MZI sensor measured at 0 and 9.15 ppbv TNT. (b) Wavelength shift of the interference dip centered at ~1550 nm with different TNT concentrations. (c) Response signals of the sensor when exposed repeatedly to TNT saturated air. The tnt/air labels represent the process of switching between TNT saturated air and dry air.
To evaluate the sensor response time, we further interrogated the wavelength shifts with 0.05 nm resolution of the OSA setting under the successive exposure of the sensor to TNT saturated air. After TNT saturated air flushing the sensor, TNT has been introduced into the air channels of PCFs for a few minutes, and subsequently the sensor has been renewed flushed with dry air for decontamination. With TNT adsorbed to the polyallylamine film, a change in the effective refractive index occurs which is measured as a wavelength shift of the interference dip. And the response signal could reach stable value. When the sensor is flushed with the dry air for desorption, a return of the signal to the initial position has been observed. This measurement has been repeated for several times in order to prove the reproducibility of the wavelength shift of the interference dip. Figure 7(c) shows the positions of the interference dip after flushing with dry air and after enrichment with TNT molecules. From the figure, we can see that the device absorbs TNT molecules rapidly but it desorbs them slowly. The response time and the recovery time (t90) were found to be nearly 90 s and 150 s, respectively. While the response has fast response time, it has the tailing effect due to the strong interaction between the polyallylamine coating and the TNT molecule. Thus, the recovery times are always greater than the response times.
But it should be noted that the LMA type interferometric sensor also has its shortage. The transfer of fluids across air channel in the PCF is critically important to film coating and gas transport. Because of intrinsic narrow gas channel in the LMA PCF the film coating uniformity is hard to be guaranteed. Not only that, but keeping the gas transfer inside is also not easy for on-site testing due to its small airway. Thus, lager gas holes in the grapefruit PCF will make it easier to coat a polymer film on the inner surface in the PCF and make it quicker to transmit gas, which might induce a better practicability than the LMA PCF in sensing applications.
3.3 Grapefruit PCF–MZI sensor response to TNT vapor
Figure 8(a) shows the transmission spectra of the grapefruit PCF–MZI sensor under exposure to dry air and TNT saturated air, respectively. The spectrum was blue shifted with increasing of the TNT concentrations. Calibration curve at ~1553 nm was linear as demonstrated in Fig. 8(b) and using the linear regression fit a sensitivity of 84 pm/ppbv can be achieved, slightly below that of the LMA type. Due to a lower Q-factor of 390, a detection limit of 1.0 ppbv can be achieved by using the calculated sensor resolution of 85 pm (3σ). Continuous response to TNT vapor shown in Fig. 8(c) exhibits very fast process of getting equilibrium under exposure to TNT. The sensor can also desorb the TNT molecule under air flushing, which demonstrates a good reversibility. The response time of the sensor was estimated at close to 60 s and the recovery time was nearly 150 s.
Fig. 8 (a) Normalized transmission spectra (arbitrary unit) of the grapefruit PCF–MZI sensor measured at 0 and 9.15 ppbv TNT. (b) Wavelength shift of the interference dip centered at ~1553 nm with different TNT concentrations. (c) Response signals of the sensor when exposed repeatedly to TNT saturated air. The tnt/air labels represent the process of switching between TNT saturated air and dry air.
3.4 Interferences with other substances
The selectivity of the LMA type and grapefruit type PCF–MZI sensor has been investigated by providing gas flows containing different substances including explosives and possible interferences. The wavelength shift of the transmission dips around 1550 nm has been measured for different analytes at the saturated vapor pressure. An approximation of sensitivity of the sensor to each analyte can be calculated and summarized in Table 1 and the magnitude of response against the vapor pressure is displayed in Fig. 9. Note that the sensor shows high degree of sensitivity to TNT and almost no response to certain polar solvents such as ethanol, acetonitrile, and acetone. However, a low sensitivity to 2,4-dinitrotoluene (DNT) can be observed owing to the similar structure to TNT. Thus, the response to other nitroaromatic need to be further studied.
Table 1. Equilibrium vapor pressure of test substances and the respective sensitivity.
3.5 Temperature characteristics of PCF–MZI sensors
Because PAH film cannot withstand high temperature, it is important to determine temperature dependence of the PCF–MZI sensors. Figure 10 shows thermal induced shift of interference dips around 1550 nm of both types of interferometric sensors. The results indicate that very low temperature dependency of 2.3 pm/°C and 3.9 pm/°C were measured for the grapefruit type and LMA type, respectively, which exhibits excellent thermostability. The heat stability of PAH coating in PCF might be mainly responsible for the thermal stability of interferometric sensor.
Fig. 10 Wavelength shift of PCF–MZI sensors as a function of the temperature change from 20 to 100°C.
3.6 Discussion
The particular structure of photonic crystal fiber in-line Mach-Zehnder interferometer sensor has many advantages over existing gas/vapor sensors. When a mixture of gas/vapor molecules transmits along the PCFs air-channels, they are separated by their specific interaction with the interior coating in the PCFs, which is extremely useful to absorb low-volatility explosives. Since the air-channel in PCFs is a few microns to dozens of micron in diameter and a few centimeters in length, it requires a very small gaseous analyte volume. As mentioned above, both types of PCF interferometer (LMA type vs. grapefruit type) have their advantages and disadvantages. The LMA type sensor possesses a larger Q-factor and a higher sensitivity to TNT concentration, while the grapefruit type fiber has such big air channels that coating inside of the fiber is quick and easy to make.
The interferometric sensor is acutely sensitive to the change in refractive index and the thickness of the interior coating. The effective index of the cladding mode will show a tiny variation depending on the capturing of the gas molecule by the coating, which leads to a shift in the interference fringes. The wavelength of the zth order interference dip can be derived as
The differential effective mode index ΔNfilm is affected by the refractive index nfilm and the thickness Thfilm of the interior coating. Considering the limited mass flux of TNT to the sensor, the change in effective index cannot be attributed to a simple change in mass addition increasing the aggregate refractive index of the film [5]. Thus, some possible amplification effect upon TNT binding must be responsible.One possible reason could be a chemical reaction that occurs upon binding of TNT. The PAH coating with a large amount of amine groups that may be capable of proton transfer with TNT [21]. Primary amines have been shown to add to the aromatic ring via creating highly colored Meisenheimer complexes [5,22,23]. But there is another possible interpretation. TNT binding may induce changes to the morphology of the film. The following presents a simple model based on the Langmuir adsorption isotherm [28]. For the PAH coating, there is a limited number of available binding sites (amine groups). Suppose that, after equilibrium is established, a fraction θ of the binding sites is occupied by adsorbed TNT molecules; a fraction 1-θ will not be occupied. The equilibrium constant can be written as K; then
so that at very low concentrationThe shift of the interference fringe depends strongly on the coating index nfilm. Thus, the interference dip shift due to the chemical interaction between the sensing film and the TNT can be presented as: , where Sn describes the sensitivity against the variations of the refractive index of film.In the coming research, more interferometric sensors based other types of microstructure fibers, such as suspended ring-core photonic crystal fiber [29] and photonic bandgap fiber [30], etc., will be explored and measured for improving gas sensing applications.
4. Conclusion
An in-line Mach-Zehnder interferometric gas sensor for explosive detection was proposed and experimentally demonstrated using the large-mode-area photonic crystal fiber and the grapefruit photonic crystal fiber, respectively. The interferometric response could be used to quantify TNT vapor over the range of 0~9.15 ppbv with detection limit down to 0.2 ppbv and has been successfully applied for the testing of TNT in situ. The sensor exhibited excellent selectivity and thermostability. This method has great potential for applications in explosive detection.
Acknowledgments
This research is supported by the National Natural Science Foundation of China (nos. 51304260 and 51574054), the Natural Science Foundation Project of CQ CSTC (no. cstc2012jjA40057), and the Scientific and Technological Research Program of Chongqing Municipal Education Commission (no. KJ1500914). The work was performed in part at the Center for Quality Engineering and Failure Prevention at Northwestern University. The authors would like to thank S. Krishnaswamy and Y. Zhu for technical assistance.
References and links
1. J. S. Caygill, F. Davis, and S. P. J. Higson, “Current trends in explosive detection techniques,” Talanta 88, 14–29 (2012). [CrossRef] [PubMed]
2. A. Hakonen, P. O. Andersson, M. Stenbæk Schmidt, T. Rindzevicius, and M. Käll, “Explosive and chemical threat detection by surface-enhanced Raman scattering: A review,” Anal. Chim. Acta 893, 1–13 (2015). [CrossRef] [PubMed]
3. Q. Zhang, D. Zhang, Y. Lu, Y. Yao, S. Li, and Q. Liu, “Graphene oxide-based optical biosensor functionalized with peptides for explosive detection,” Biosens. Bioelectron. 68, 494–499 (2015). [CrossRef] [PubMed]
4. R. A. Mullis, A. L. Witzel, and J. Price, “Maintenance energy requirements of odor detection, explosive detection and human detection working dogs,” PeerJ 3, e767 (2015). [CrossRef] [PubMed]
5. P. L. Edmiston, D. P. Campbell, D. S. Gottfried, J. Baughman, and M. M. Timmers, “Detection of vapor phase trinitrotoluene in the parts-per-trillion range using waveguide interferometry,” Sens. Actuators B Chem. 143(2), 574–582 (2010). [CrossRef]
6. Y. Zhao, X. Li, L. Cai, and Y. Yang, “Refractive index sensing based on photonic crystal fiber interferometer structure with up-tapered joints,” Sens. Actuators B Chem. 221, 406–410 (2015). [CrossRef]
7. B. Kim, T.-H. Kim, L. Cui, and Y. Chung, “Twin core photonic crystal fiber for in-line Mach-Zehnder interferometric sensing applications,” Opt. Express 17(18), 15502–15507 (2009). [CrossRef] [PubMed]
8. S. Krishnaswamy, S. Zheng, and Y. Zhu, “Photonic crystal fiber long-period gratings as a fiber-optic sensing platform for applications in structural health monitoring,” in Proc. 6th European Workshop on Structural Health Monitoring (2012), paper Tu.4.C.3.
9. D. Wu, Y. Zhao, and J. Li, “PCF taper-based Mach–Zehnder interferometer for refractive index sensing in a PDMS detection cell,” Sens. Actuators B Chem. 213, 1–4 (2015). [CrossRef]
10. B. Wang, W. Zhang, J. Li, T. Yan, L. Zhang, L. Chen, and Q. Zhou, “Mach–Zehnder interferometer based on S-tapered all-solid photonic bandgap fiber,” IEEE Photonics Technol. Lett. 27(17), 1849–1852 (2015). [CrossRef]
11. W. Huang, Y. G. Liu, Z. Wang, B. Liu, J. Guo, M. Luo, and L. Lin, “Intermodal interferometer with low insertion loss and high extinction ratio composed of a slight offset point and a matching long period grating in two-mode photonic crystal fiber,” Appl. Opt. 54(2), 285–290 (2015). [CrossRef] [PubMed]
12. Y. Yu, Y. Zhang, Z. Ou, X. Chen, Q. Huang, and S. Ruan, “Simultaneous measurement of one dimensional bending and temperature based on Mach-Zehnder interferometer,” Photonic Sens. 5(4), 376–384 (2015). [CrossRef]
13. J. H. Liou and C. P. Yu, “All-fiber Mach-Zehnder interferometer based on two liquid infiltrations in a photonic crystal fiber,” Opt. Express 23(5), 6946–6951 (2015). [CrossRef] [PubMed]
14. X. Bai, D. Fan, S. Wang, S. Pu, and X. Zeng, “Strain sensor based on fiber ring cavity laser with photonic crystal fiber in-line Mach–Zehnder interferometer,” IEEE Photonics J. 6(4), 6801608 (2014).
15. J. Villatoro, V. Finazzi, G. Badenes, and V. Pruneri, “Highly sensitive sensors based on photonic crystal fiber modal interferometers,” J. Sens. 2009, 747803 (2009). [CrossRef]
16. W. Shin, Y. L. Lee, B.-A. Yu, Y.-C. Noh, and T. J. Ahn, “Highly sensitive strain and bending sensor based on in-line fiber Mach–Zehnder interferometer in solid core large mode area photonic crystal fiber,” Opt. Commun. 283(10), 2097–2101 (2010). [CrossRef]
17. X. Chen, Y. Yu, X. Xu, Q. Huang, Z. Ou, J. Wang, P. Yan, and C. Du, “Temperature insensitive bending sensor based on in-Line Mach-Zehnder interferometer,” Photonic Sens. 4(3), 193–197 (2014). [CrossRef]
18. S. H. Kassani, J. Park, Y. Jung, J. Kobelke, and K. Oh, “Fast response in-line gas sensor using C-type fiber and Ge-doped ring defect photonic crystal fiber,” Opt. Express 21(12), 14074–14083 (2013). [CrossRef] [PubMed]
19. J. Harris, P. Lu, H. Larocque, L. Chen, and X. Bao, “In-fiber Mach–Zehnder interferometric refractive index sensors with guided and leaky modes,” Sens. Actuators B Chem. 206, 246–251 (2015). [CrossRef]
20. H. Y. Choi, M. J. Kim, and B. H. Lee, “All-fiber Mach-Zehnder type interferometers formed in photonic crystal fiber,” Opt. Express 15(9), 5711–5720 (2007). [CrossRef] [PubMed]
21. L.-Q. Lu, Y. Zheng, W.-G. Qu, and A.-W. Xu, “Convenient and sensitive synchronous fluorescence detection of trace TNT based on FRET using FITC–PAH as a probe,” Anal. Methods 5(3), 603–607 (2013). [CrossRef]
22. R. Ban, F. Zheng, and J. Zhang, “A highly sensitive fluorescence assay for 2,4,6-trinitrotoluene using amine-capped silicon quantum dots as a probe,” Anal. Methods 7(5), 1732–1737 (2015). [CrossRef]
23. X. Zhou, H. Liu, L. Yang, and J. Liu, “SERS and OWGS detection of dynamic trapping molecular TNT based on a functional self-assembly Au monolayer film,” Analyst (Lond.) 138(6), 1858–1864 (2013). [CrossRef] [PubMed]
24. R. G. Ewing, M. J. Waltman, D. A. Atkinson, J. W. Grate, and P. J. Hotchkiss, “The vapor pressures of explosives,” Trends Analyt. Chem. 42, 35–48 (2013). [CrossRef]
25. I. M. White and X. Fan, “On the performance quantification of resonant refractive index sensors,” Opt. Express 16(2), 1020–1028 (2008). [CrossRef] [PubMed]
26. A. Freedman, P. L. Kebabian, Z. Li, W. A. Robinson, and J. C. Wormhoudt, “Apparatus for determination of vapor pressures at ambient temperatures employing a Knudsen effusion cell and quartz crystal microbalance,” Meas. Sci. Technol. 19(12), 125102 (2008). [CrossRef]
27. Dortmund Data Bank, http://ddbonline.ddbst.de/AntoineCalculation/AntoineCalculationCGI.exe.
28. J. Yang, C. Tao, X. Li, G. Zhu, and W. Chen, “Long-period fiber grating sensor with a styrene-acrylonitrile nano-film incorporating cryptophane A for methane detection,” Opt. Express 19(15), 14696–14706 (2011). [CrossRef] [PubMed]
29. S. H. Kassani, R. Khazaeinezhad, Y. Jung, J. Kobelke, and K. Oh, “Suspended ring-core photonic crystal fiber gas sensor with high sensitivity and fast response,” IEEE Photonics J. 7(1), 1–9 (2015). [CrossRef]
30. J. Sun, C.-C. Chan, X.-Y. Dong, and P. Shum, “High-resolution photonic bandgap fiber-based biochemical sensor,” J. Biomed. Opt. 12(4), 044022 (2007). [CrossRef] [PubMed]
