The feasibility of remote chemical detection is experimentally demonstrated by using a Teflon pipe as a scanning arm in a continuous-terahertz wave sensing and imaging system. Different tablets with distinct mixed ratios of aluminum and polyethylene powders are well distinguished by measuring the power reflectivities of 0.4 THz wave associated with their distinct terahertz refractive indices. Given its refractive index sensitivity and fast response, the reflective terahertz sensing system can be used to real-time trace and quantitatively analyze the ammonium–chloride aerosols produced by the chemical reaction between hydrochloric acid and ammonia vapors. With a tightly focusing terahertz beam spot, the spatial and concentration distributions of the generated chemical product are successfully mapped out by the 1D scan of the flexible pipe probe. In consideration of the responsitivity, power stability, and focused spot size of the system, its detection limit for the ammonium–chloride aerosol is estimated to be approximately 165 nmol/mm2. The reliable and compact terahertz pipe scan system is potentially suitable for practical applications, such as biomedical or industrial fiber endoscopy.
© 2016 Optical Society of America
The development of fiber sensors for the detection of chemical or biological substances has attracted considerable attention over the last decade [1, 2]. The general advantages of optical fibers include portability, flexibility, compatibility with various electromagnetic fields, and easy integration with other sensing units, such as microfluidic channels, laboratory chips, and specific waveguides. Optical fibers can readily deliver optical signals to an inaccessible region for the remote and real-time tracing of hostile and hazardous agents without the difficult collection process. Optical fibers may be used to detect explosive/flammable gases in mines or industrial sites , to trace the adsorbents on the inner surface of industrial pipelines , monitor environmental pollutants , and inspect human breath , among others.
Terahertz (THz) radiation with a frequency range of 0.1–10 THz has emerged as a promising new tool for chemical imaging and sensing applications  based on its unique properties. Many nonmetallic and nonpolar materials are transparent in the THz frequency range but opaque to the visible/infrared-ray(IR) light. Moreover, the penetration depth of certain THz waves in the dry chemical substances is deeper than that of optical light because of the long wavelength and less scattering of THz waves. Therefore, compared with optical waves, THz waves have a stronger potential for the label-free and nondestructive identification of concealed or buried objects through remote detection. Current THz imaging systems are either fiber-based  or free-space systems that are constructed by bulky reflectors and a sample-scanning unit for broad-range detection . Especially, the fiber-scanned modality can meet the requirements for the remote detection of hidden targets at any location and the prevention of delivered signals from environmental disturbance. The fiber-scanning system has the advantages of compact size, portability, and high flexibility. This system has been extensively employed in imaging and sensing applications at the IR and visible spectral ranges [10, 11].
Two kinds of THz fiber-based imaging/sensing systems have been successfully demonstrated for biomedical diagnosis and mapping out the surface structure or water distributions of bio-specimens [12–14]. One system is based on the 2D scan of a tiny probe composed of THz photoconductive switch which is connected to an optical fiber . However, this sensing modality should be operated with a high voltage (i.e., near 100 V) for the photoconductive switch, and there is dangerous concern to in vivo detect biotissues or the gaseous explosives. In addition, the photoconductive-switch-scan system requires a time-delay-scanning unit to obtain a complete time–domain waveform that carries information about the analyte. Thus, this system is unsuitable for real-time response and in situ detection. The second method is based on the scan of a low-loss subwavelength THz fiber [8, 13]. It is comparably safer than the photoconductive probe scan method and can be conjugated with the electronic THz system  to real-time detect the reflected THz wave power from the sample surface, thereby avoiding the need for a time-delay-scanning unit. The electronic THz emitter/detector features a fast response and easy operation; it is also highly suitable for real-time and in situ imaging/sensing applications. However, the high bending loss of the subwavelength fiber restricts its practical applications because most of the waveguide power is distributed in the air cladding [8, 13, 14]. The subwavelength fiber-based THz system also encounters difficulties in flexibility and portability. To date, the development of THz sensing/imaging based on the fiber scanning method is still in its infant stage, and its practicability in remote chemical sensing has yet to be explored.
In this study, we successfully demonstrate the feasibility of in situ and remote chemical detection through the use of a reflective THz fiber-scan imaging system to trace the dynamic generation of a compound during a chemical reaction and concurrently to map out the spatial distribution of the chemical product. The reflective THz fiber-scan system is demonstrated at 0.4 THz emitted from a continuous THz wave source, and a simple Teflon pipe with low THz transmitting and bending loss [16–18] is used as a sensing and image scanning arm. The bending capability of the reflective THz–pipe–scan system is experimentally characterized and proven by identifying a tablet array through its pipe-scanned THz image. The chemical sensing/imaging capability of the system is validated through remote measurement of the reflected THz power from a chemical product, ammonium chloride (NH4Cl) aerosols, which is dynamically generated by the reaction between hydrochloric acid (HCl) and ammonia (NH3) vapors. The sensing mechanism and detection sensitivity of the chemical product is also discussed in detail.
2. Configuration and sensitivity calibration of a THz pipe scan system
The configuration of a THz pipe scan system is shown in Fig. 1(a). The system involves two parts: an electrical THz transmitting/receiving unit and a fiber scan unit. The signal from the Gunn oscillator passes through a frequency multiplier and feeds into a horn antenna for continuous-THz wave radiation. The adopted Gunn oscillator module provides THz power at 0.4 THz around 100 µW. The free-running THz waves are collimated and focused into a Teflon pipe fiber by a pair of off-axis parabolic mirrors (50 mm effective focal length). The flexible Teflon pipe has a length of 30 cm, an inner hollow core diameter of 8.5 mm, and a pipe wall thickness of 1 mm. The pipe is mounted on a 2D motional stage as the scanning fiber of the system. The former 6 cm length of the pipe apart from the input end is fixed for well input coupling, and the remaining 24 cm-long pipe is bendable for 2D scan. THz waves are delivered in the plastic pipe on the basis of the antiresonant reflecting waveguiding principle , and the single-mode propagation in the hollow core causes low loss and low dispersion [17, 18]. A plastic ball lens that is composed of polyethylene (PE) material has a 10mm-diameter and is cut into two hemispherical lenses by a knife. The polished PE hemispherical lens has a 5 mm-effective focal length and is glued at the output end of the pipe to focus the transmitted THz waves onto the sample surface. The focused THz spot size after the plastic hemispherical lens is around 1.4 mm, measured by the knife-edge method and determined from the 10-90% power transmission. A gold mirror is placed approximately 5 mm apart from the apex of the PE hemispherical lens to obtain the optimized THz reflected power. The reflective THz radiations from the gold mirror are then passed back through the same pipe, reflected by a beam splitter (composed of a 1 mm-thick silicon plate), and then coupled into a THz detector by a plastic lens for power detection. The used THz detector is a zero bias Schottky diode with a responsivity of 1500 V/W at 0.4 THz, provided from Virginia Diode Inc. To efficiently couple the THz radiation from the free space into the Schottky diode, a horn antenna is assembled at the input aperature of the Schottky diode. A chopper and a lock-in amplifier are used in the system to raise the signal-to-noise ratio (SNR) approaching 200. In this experiment, the measured THz power is represented as the output voltage of the Schottky diode instead of watt. It should be noted that the detected THz power cannot be directly obtained from the responsivity of the Schottky diode beacuse the efficiency of the horn antenna and the lock-in amplifier are not calibrated in the provided responsivity.
To investigate the bending loss of the Teflon pipe, a gold mirror is assembled with the pipe 5 mm apart from the fiber output end to maintain the normal incidence and the same focused position at various bending angles. By holding the pipe as straight about 6 cm apart from the input end and laterally moving the pipe output end, the measured bending loss of the THz pipe as a function of bending angle and radius can be obtained [Fig. 1(b)]. The reflectivity, which is defined by the powers (Pθ) at the bending angle (θ) and that without any bending (P0), can be expressed as “Pθ/P0”, and the corresponded bending loss equals “1-(Pθ/P0).” The bending loss curves are almost identical in both clockwise and counterclockwise bending directions, and the performance of the bending loss is quite similar to the theoretical result demonstrated in the reference . The bending loss is less than 10% within a +/−5 degree bending angle and can be neglected within a +/−1 degree bending angle, which is much smaller than that of the subwavelength plastic wire .
In this experiment, 12 mm-diameter tablets serve as standard samples to assess the sensitivity for the sample’s optical properties by using the pipe scan reflective THz imaging system. Various volume ratios of aluminum (Al) (Shimakyu Pure Chemical Co., grain size ~40 μm) and PE powders (Shimakyu Pure Chemical Co., grain size ~60 μm) are mixed and compressed into tablets, and the mixed Al/PE ratios include 10/0, 7/3, 5/5, 3/7, and 0/10. Figure 1(c) shows the schematic of remotely scanning a fastened tablet using the pipe fiber, where the tablet is placed at the focus of the hemispherical PE lens. There are THz standing waves between the tablet and the apex of the PE hemispherical lens due to the long coherent length of the THz-continuous wave. At different reflector positions along the optic axis, the detected THz power apparently varies while scanning the THz fiber . However, the axial power variation can be normalized by comparing the measured THz reflection power of the tablet to that of a metal mirror, and consequently prevented in the THz image. Figures 1(d) and 1(e) show the visible and THz wave images of a tablet array with different Al/PE mixing ratios, respectively. The distinct color intensity (normalized to 1) of each tablet image in Fig. 1(e) represents different THz normalized reflection power. The bending angle of a tablet image (12 mm in diameter) is within +/−1 degree. Therefore, the reflected power difference due to the fiber bending can be neglected. Combing the negligible lateral bending loss and the calibrated axial power variation, the intensity uniformity of the reflected THz waves can be well preserved, and the sample information from different tablets would not distort during fiber scanning as shown in Fig. 1(c).
Figure 1(e) shows that the difference in the Al/PE volume ratio can be distinguished on the basis of the reflected THz power. We extracted the THz reflectivity at zero bending angle from the measured results of Fig. 1(e) and plotted it relative to the alimunum-volume ratio of each tablet in Fig. 2. Figure 2(a) shows the measured and calculated reflectivities of various tablets with different Al powder percentages. The calculated reflectivities are based on the Fresnel formula [20, 21] under normal incidence and defined as (neff − nair)2/(neff + nair)2, where neff and nair are respectively the effective refractive indices of a tablet and air. The theoretical THz refractive indices of the tablets can be estimated on the basis of effective medium theory by using the equation neff = (nAl ∙ σAl) + (nPE ∙ σPE). In this equation, nAl and nPE are the refractive indices of Al and PE bulk materials at 0.4 THz (nAl = 29.64  and nPE = 1.55 ), respectively, and σAl and σPE are the volume ratios of Al and PE powders in a tablet, respectively. From the analysis of Fig. 2(a), the measured THz refractive indices of different tablets can also be obtained via the Fresnel formula [Fig. 2(b)]. The measured reflectivity and refractive indices at 0.4 THz are quite consistent with the theoretical results shown in Figs. 2(a) and 2(b), respectively, and the THz refractive index is approximately proportional to the fractional volume of Al powder. Figure 2 indicates that using the pipe scan THz system can indeed distinguish the distinct THz-refractive indices of various samples based on the Fresnel reflectivity of 0.4 THz wave. In this experiment, the effective THz refractive index of tablet is measured by the focused THz beam in a reflection system and a slight variation in THz refractive index due to the Gouy phase shift  can be negligible because the skin depth of THz beam in the metal-dielectric mixed tablet is smaller than THz wavelength . The Gouy shift inducing refractive index error merely occurs in a very large sample thickness, and the measured index deviation using a transmission THz time-domain spectroscopy is around 0.23% comparing to the measurement by a collimating THz beam .
The sensitivity to the refractive index detection using the THz pipe scan system can be evaluated from Fig. 2(a) on the basis of the tangential slopes of the curve at different fractional volumes of Al. The fractional volume of Al can be translated into the THz refractive indices of the samples [Fig. 2(b)]. The relation of tangential slopes to the different THz refractive indices of the samples is plotted to specify the sensitivity performance for any object with a certain THz refractive index. Figure 3 illustrates the sensitivity of THz refractive index sensing. The optimized performance in this experiment is to identify a tablet within the refractive index range of 1–3. In other words, the sensitivity of the system decreases when detecting a sample with a large THz refractive index (i.e., > 3.0). This result indicates that the pipe scan system has a high sensitivity for detecting the THz refractive index variation of low-density objects because such materials as vaporized molecules or aerosols generally have low THz refractive indices .
3. Chemical reaction sensing based on a THz pipe scan system
In this section, we employ the highly sensitive pipe scan system for the remote and in situ detection of aerosol generation during a chemical reaction. The aerosol in this experiment is the chemical product of NH4Cl, which is generated by the vapor reactants of the HCl and NH3 molecules. Figure 4(a) shows the optical photo of a tubular chemical reaction chamber composed of a polypropylene (PP) tube, 20 cm in length, 9 mm in outer diameter, and 0.25 mm in tube wall thickness. The PP tubular chamber provides a sufficient air space and a transparent tube wall to observe the generation of NH4Cl aerosol with the naked eyes. The two ends of the tubular chamber are sealed by a pair of PE caps to avoid gas leakage. Two flexible hoses are used to fasten two PE caps on the PP chamber and assist liquid injection without moving the fluidic system. As shown in Fig. 4(b), a cotton sorbent is placed inside the PE cap to absorb the liquid reactants and jointed a analyte inlet of Teflon microtube. A pair of Teflon microtubes is pierced through the holes on the PE cap for the injection of chemical reactants [Fig. 4(c)]. One of the inlets is used to input HCl liquid (Wako Pure Chemical Industries, Ltd., 500mL, 35%), and the other is used to input NH3 liquid (Wako Pure Chemical Industries, Ltd., 500mL, 28%). In consideration of the almost identical weight of the two cotton sorbents, the two liquid reactants are assumed to evaporate from the same interface area of the moist cotton. The injected liquid volumes of the HCl and NH3 molecules are identical and manipulated by the syringes, which are connected to the Teflon microtubes. The reactant liquids are injected into the tubular chamber under ambient room temperature and normal atmosphere pressure without vacuum pumping. After injecting the liquids, the syringes are removed and then the microtubes are sealed. Figure 4(d) shows the configuration of the scanning pipe arm and the tubular chamber for the generation of NH4Cl aerosol. The reactant liquids of HCl and NH3 are concurrently injected, becoming vapors from the opposite ends of the tubular chamber, and then diffuse toward the chamber middle region until the two reactant vapors meet. Then, NH4Cl aerosol is produced, sedimented, and physically adsorbed on the chamber wall while the vaporized molecules of HCl and NH3 interact under the saturated vapor pressure of the mixed reactant vapors. The white adsorbed aerosol can extend around 60 mm length of the reaction chamber and be directly observed by the naked eyes. Only the central 30 mm length is scanned by the Teflon pipe for sensing because of the limited THz wave power at large bending angles.
To detect the NH4Cl aerosol inside the tubular chamber, the 0.4 THz wave is approximately focused onto the inner surface of the chamber wall when the apex of the hemispherical PE lens is separated approximately 5 mm from the outer surface of the PP chamber. Figure 5(a) shows the dynamic relative reflectivity of the 0.4 THz wave within the first 15 min of the chemical reaction. The THz reflected power is recorded by THz detector with a single acquisition of the power per second, and the scan pipe is fixed and remained straight for the delivery and collection of THz waves reflected from the chemical molecules. The dynamic THz response from chemical molecules can be in situ detected because the liquid injection and THz power measurement start at the same time. The relative reflectivity in Fig. 5(a) represents the power ratio between the reflected power recorded during the dynamical chemical reaction and that at 0 min (i.e., the blank PP tube without any analyte). The maximum noise deviation level of THz reflected power is less than 0.2µV measured by the reflective system with a SNR of 200, corresponding to < 0.5% variation of the relative reflectivity. Based on the relative reflectivity variation, the three molecules of HCl, NH3, and NH4Cl interacting with the 0.4 THz wave are consequently analyzed. Different reactant volumes (HCl and NH3 liquids) of 0.03, 0.06, 0.10, and 0.15 cm3 are used to generate different amounts of the chemical product NH4Cl and dynamically detected by the Teflon pipe system. As shown in Fig. 5(a), the 0.4 THz wave is apparently absorbed by the HCl and NH3 vapors within the initial 3 min; so that, the relative reflectivity is less than 1. In other words, the power reflected from the HCl and NH3 vapors is smaller than that of the air in a blank chamber for all injection volumes because of vapor absorption. The variations of the relative reflectivities from HCl and NH3 vapor absorptions can thus be reliably recognized using the system with noise level smaller than 0.5%-relative reflectivity variation. After 3 min of reaction, the relative reflectivity is gradually increased as the NH4Cl aerosol is generated. The THz reflectivity related to NH4Cl aerosol generation is steadily increased until equilibrium within 3–15 min. The time periods of vapor absorption for 0.03, 0.06, 0.10, and 0.15 cm3 injection volumes are 1–189, 1–259, 1–205, and 1–146 s, respectively. The corresponding equilibrium periods, that is, when the THz relative reflectivity reaches steady state, for the 0.03, 0.06, 0.10, and 0.15 cm3 injection volumes are 347–900, 399–900, 411–900, and 287–900 s, respectively. The steady THz reflectivity within the equilibrium period is mainly contributed from the adsorbed NH4Cl aerosol because the suspended NH4Cl aerosols almost sediment and physically adsorb on the tube wall at the later equilibrium period (after 8 min) for all injection volumes. The average THz reflectivities within the periods of vapor absorption and equilibrium states for all injection volumes are obtained and statistically analyzed in Fig. 5(b). The average relative reflectivity in the equilibrium state is highly correlated with reactant volumes with a linearly proportional relation, but the THz relative reflectivity in the vapor absorption period is independent of the reactant volumes. This result indicates that the generated NH4Cl aerosol has a higher THz refractive index than the air space and that the high index apparently leads to considerably larger THz wave reflection than THz wave absorption [Fig. 5(a)].
To quantitatively investigate and map out the distribution of the chemical product using the pipe scan system, the Teflon pipe is used to 1D scan the NH4Cl-adsorbed chamber after the liquid injection about 8 minutes meanwhile the THz relative reflectivity is almost constant. Figure 6(a) shows the aerosol generation based on the 0.03 cm3 HCl and NH3. The white region, distributed at the range of 11–13 mm, represents the physically adsorbed NH4Cl aerosol and is marked by a ruler. The transverse length of fiber scan is approximately 30 mm of the PP chamber. Figure 6(b) shows the THz reflected power within the 30 mm scan length, which is acquired before and after the chemical reaction and respectively shown as black and red lines. The red-filled area in Fig. 6(b) reflects the power difference between the black and red lines, and its distribution agrees with the white region location (i.e., 11–13 mm) shown in Fig. 6(a). In Fig. 6(b), the maximum THz power difference is approximately 1.0 µV and the maximum noise deviation level of THz reflected power is less than 0.2µV. Hence, the measured THz power difference from the chemical analyte is sufficiently larger than the system noise level for recognition. On the basis of the same configuration, the distribution of the NH4Cl aerosol for the more reactant volume 0.06 cm3 is also mapped out by the THz fiber scan system and shown in Figs. 6(c) and 6(d). The occupied length of the white region is obviously greater compared with that of the 0.03 cm3 condition and located at 5–15 mm [Fig. 6(c)]. On the basis of the sensing results illustrated in Fig. 6(d), the power increment contributed by the aerosols matches with the white region location in the optical photo [Fig. 6(c)], and the maximum power difference is approximately 1.8 µV. The low reflected THzpower in the 0–5 and 25–30 mm regions is attributed to the high bending loss of the Teflon pipe. This phenomenon leads to a high uncertainty of the reflected power values, which is not suitable to identify the chemical analytes. In other words, the detected THz power and power-difference response contributed by the adsorbed chemical molecules is not repeatedly measured at the scan edge areas. Figures 6(e) and 6(f) show the sensing results of the NH4Cl aerosol generation for the injected reactant volumes of 0.10 and 0.15 cm3, respectively. The maximum power differences for the 0.10 and 0.15 cm3 conditions are 1.8 and 2.6 µV, respectively. A comparison of the white region locations recorded in the optical photos with the red-filled areas of the THz-wave scanned results in Fig. 6 reveals that the aerosol distribution area can be well identified for different injection volumes using the THz pipe scan system. Additionally, the peak response of power difference, as illustrated in Figs. 6(b), 6(d)–6(f), is all located at around 12mm position near the left (HCl) side of the tubular chamber because the diffusion speed of the HCl vapor is lower than that of the NH3.
The aerosol distribution in the tubular chamber is generally nonuniform. Thus, quantitatively analyzing the chemical product at a specific spatial position is inadequate. Therefore, the total THz signals reflected from the white region sections (i.e., 5–15 mm) are summed for the integral of the THz power difference with respect to the pipe scan position. It corresponds with the red-filled areas in Figs. 6(b), 6(d)–6(f) and summarized in Fig. 7(a). The maximum values of the integrals for the different injection volumes are distinct and represent the summation of reflective THz power from the generated NH4Cl aerosol along the individual scans. The maximum integral values in Fig. 7(a) for the 0.03, 0.06, 0.10, and 0.15 cm3 injection conditions are 2, 7, 8, and 9 µV·mm, respectively. The maximum integral value apparently increases with the reactant volume and is consistent with the proportional relation between the peak power difference and the reactant volume as observed in Fig. 6. Figures 7(b) and 7(c) illustrate the total amounts of the generated NH4Cl aerosol in the tubular chamber and the corresponding average THz power differences under different injected reactant volumes, respectively. The weight of the NH4Cl aerosol is measured by a precise electronic balance and obtained by comparison of the chamber weights before and after the chemical reaction. The average THz power differences are acquired from Figs. 6(b), 6(d)–6(f). For the injected volumes of 0.03, 0.06, 0.10, and 0.15 cm3, the generated aerosol weights are 0.66, 0.74, 1.015, and 1.16 mg, respectively, and the corresponding average values of the THz power-difference are 0.71, 1.30, 1.42, and 2.07 µV. The error bar in Fig. 7(c) represents the measured power variation in each scan process. The manifestations in the generated weights of the NH4Cl aerosol [Fig. 7(b)] and the average THz power difference [Fig. 7(c)] almost have the same trend, and both are approximately proportional to the injected reactant volumes. This result is similar to that illustrated in Fig. 5(b), which are directly correlated with the generated amounts of NH4Cl aerosol. According to the proportional relation between the THz wave power difference and the aerosol amount, we assume that the power difference is related to the thickness of the aerosol layer adsorbed on the inner chamber wall. The aerosol amount detected within a THz wave beam spot (~1.4 mm diameter) can therefore be estimated using the equationFigure 7(d) illustrates the estimated Wspot under different injected reactant volumes and presents almost a proportional relation. The detected chemical product weight inside a THz wave beam spot for the 0.03, 0.06, 0.10, and 0.15 cm3 reactant injections are 37, 42, 58, and 66 µg, respectively. Combining the results of Figs. 7(c) and 7(d), we can plot the relation of the THz response to the detectable product weight within one THz spot and linear fit in Fig. 7(e). On the basis of the average deviation of the measured THz response [~0.57 µV as shown in the error bar of Fig. 7(c)] and the slope of linearly fitting curve [0.0418 µV/ µg as shown in Fig. 7(e)], the detection limit of the generated NH4Cl aerosol using the pipe scan image system can be estimated to be approximately 13.63 µg, which corresponds to 165 nmol/mm2.
This study experimentally demonstrates the feasibility to in situ and dynamically monitor a chemical reaction by using a THz fiber scan system. Refractive index sensing of surface-adsorbed chemicals is conducted by measuring the power change of the reflected THz wave before and after the chemical reaction via the 1D motional scan of a low-loss 30 cm-long Teflon pipe. The chemical reaction of HCl and NH3 vapors to generate the aerosol compound NH4Cl in the sealed chamber is used as an example. The remote sensing results for the chemical distributions and the measured refractive indices of mixed tablets agree with the optical images and the theoretical results. For the quantitative analysis of the chemical product inside the reaction chamber, the amount of NH4Cl aerosol is approximately proportional to the power increment of the reflected THz wave found in both dynamically monitoring THz reflectivity and the 1D spatial scanning result. The measurement of THz reflectivity from the surface-adsorbed chemicals is repeatable and reliable on the basis of the pipe fiber scan. The sensitivity of the pipe fiber sensing system for the NH4Cl aerosol is evaluated to be approximately 13.63 µg, which corresponds to 165 nmol/mm2. The demonstrated THz pipe scan imaging/sensing modality has the advantages of flexibility, compactness, and remote operation and is potentially suitable for applications, such as homeland security, biomedical or industrial fiber endoscopy, and environmental pollutant monitoring.
This work was supported by the Advanced Optoelectronic Technology Center of National Cheng-Kung University under projects from the Ministry of Education and the Ministry of Science and Technology (MOST 104-2221-E-006-163-MY3) of Taiwan.
References and links
3. 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]
4. A. Albertazzi, A. C. Hofmann, A. V. Fantin, and J. M. C. Santos, “Development and application of a photogrammetric endoscopic system for measurement of misalignment and internal profile of welded joints in pipelines,” Proc. SPIE 7389, 73891W (2009).
5. L. S. Grattan and B. T. Meggitt, Optical Fiber Sensor Technology: Chemical and Environmental Sensing (Springer Science & Business Media, 2013), Chap. 5.
6. J. Heo, M. Rodrigues, S. J. Saggese, and G. H. Sigel Jr., “Remote fiber-optic chemical sensing using evanescent-wave interactions in chalcogenide glass fibers,” Appl. Opt. 30(27), 3944–3951 (1991). [CrossRef] [PubMed]
8. J. Y. Lu, C. M. Chiu, C. C. Kuo, C. H. Lai, H. C. Chang, Y. J. Hwang, C. L. Pan, and C. K. Sun, “Terahertz scanning imaging with a subwavelength plastic fiber,” Appl. Phys. Lett. 92(8), 084102 (2008). [CrossRef]
9. Y. Cui, W. Fu, X. Guan, M. Hu, Y. Yan, and S. Liu, “Experiment studies on two-dimension terahertz raster scan imaging,” J. Infrared Milli. Terahz. Waves 33(5), 513–521 (2012). [CrossRef]
11. C. M. Lee, C. J. Engelbrecht, T. D. Soper, F. Helmchen, and E. J. Seibel, “Scanning fiber endoscopy with highly flexible, 1 mm catheterscopes for wide-field, full-color imaging,” J. Biophotonics 3(5-6), 385–407 (2010). [CrossRef] [PubMed]
13. J. Y. Lu, C. C. Kuo, C. M. Chiu, H. W. Chen, Y. J. Hwang, C. L. Pan, and C. K. Sun, “THz interferometric imaging using subwavelength plastic fiber based THz endoscopes,” Opt. Express 16(4), 2494–2501 (2008). [CrossRef] [PubMed]
15. P. H. Siegel, “Terahertz technology,” IEEE Trans. Microw. Theory Tech. 50(3), 910–928 (2002). [CrossRef]
17. C. H. Lai, B. You, J. Y. Lu, T. A. Liu, J. L. Peng, C. K. Sun, and H. C. Chang, “Modal characteristics of antiresonant reflecting pipe waveguides for terahertz waveguiding,” Opt. Express 18(1), 309–322 (2010). [CrossRef] [PubMed]
20. N. Laman and D. Grischkowsky, “Reduced conductivity in the terahertz skin-depth layer of metals,” Appl. Phys. Lett. 90(12), 122115 (2007). [CrossRef]
21. B. E. A. Saleh and M. C. Teich, Fundamentals of Photonics (Wiley, 1991), Chap. 6.
22. M. Naftaly, “Metrology issues and solutions in THz time-domain spectroscopy: noise, errors, calibration,” IEEE Sens. J. 13(1), 8–17 (2013). [CrossRef]
23. M. A. Ordal, L. L. Long, R. J. Bell, S. E. Bell, R. R. Bell, R. W. Alexander Jr, and C. A. Ward, “Optical properties of the metals Al, Co, Cu, Au, Fe, Pb, Ni, Pd, Pt, Ag, Ti, and W in the infrared and far infrared,” Appl. Opt. 22(7), 1099–1120 (1983). [CrossRef] [PubMed]
24. J. W. Lamb, “Miscellancous data on materials for millimetre and submillimetre optics,” Int. J. Infrared. Milli. 17(12), 1996–2034 (1997). [CrossRef]