Open Access
Add to BibSonomy
Abstract
We report on a flexible platform for molecular sensing in the terahertz range. Merging the well-established technologies of near-infrared electro-optic modulation and photomixing realizes a spectrally adaptable terahertz source, which is combined with a new generation of compact gas cells, so-called substrate-integrated hollow waveguides (iHWGs). iHWGs have been developed in the mid-infrared and provide flexibility in the design of the optical absorption path. Here, we demonstrate its suitability for the terahertz domain by presenting its low propagation losses and by measuring rotational transitions of nitrious oxide (N2O). A fast frequency sideband modulation technique results in substantially reduced measurement times and increased accuracy compared to a standard wavelength tuning method.
Published by Optica Publishing Group under the terms of the Creative Commons Attribution 4.0 License. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI.
1. Introduction
Trace gas detection in the terahertz (THz) range (0.3–10 THz) has gained increased attention in recent years with the continuous development of THz sources and detectors. The huge amount of molecules relevant for environmental monitoring and medical analysis exhibit fundamental rotational transitions in that spectral region [1,2]. Some of their absorption cross-sections are orders of magnitude larger compared to those of the vibrational transitions of these molecules in the well-established NIR domain (e.g. H$_2$O and O$_3$ [3]). The narrow THz frequency window compared to the broad NIR window, also offers the potential to develop tunable devices that can target different application dependent molecules with a single measuring system. This gradually makes the THz field competitive by means of gas sensing. The key to practical applications is the development of low-cost, fast and flexible systems that remain compact and robust.
For gas sensors based on direct light absorption at specific wavelengths the interaction length of the gaseous species and the photons (i.e. the optical path length OPL) is crucial in terms of sensitivity [4]. Multipass gas cells (e.g. Herriott or White) provide an extended OPL by enabling multiple passes of the probing light beam through a gas volume inside the cell [5–7]. They are usually quite bulky, need an optical alignment and require a relatively large volume of the gas sample reaching up to liters, of which only a small fraction is probed due to the geometry necessary for multiple beam reflections. Spectroscopy based on cavity enhancement effects might provide an alternate solution [8], although its progress still faces difficulties in the demanding fabrication of high-finesse THz cavities. Another promising alternative is presented by THz optical fibers and waveguides [9,10]. There are examples where solid core fibers are used for sensing in form of interaction with the evanescent THz field [11]. Nevertheless, more attention is paid to hollow core designs serving essentially as light pipes that guide the radiation inside a hollow core of a fiber or waveguide and simultaneously establish an efficient miniaturized gas cell with predefined optical absorption path length [12,13]. Milliliters or less of the sample gas volume is sufficient to fill the hollow core and provide maximal overlap between the analyte and the guided THz light. However, conventional hollow waveguides (HWGs) often struggle when it comes to compactness and flexibility as well as suffer from mechanical vibrations and temperature fluctuations. Therefore, they require careful handling or additional protection and are usually quite demanding in production including internal coatings. It should be mentioned that the development on 3D printing techniques for hollow core THz fibers shows the potential to simplify the production process a lot [14].
Recently, Mizaikoff et al. developed a new concept that maintains the advantages of conventional HWGs and resolves their major drawbacks [15]. The so called substrate-integrated hollow waveguide (iHWG) represents a miniaturized gas cell assembled by a solid substrate material (i.e. aluminium) with integrated straight or meandered reflective light-guiding channels and a sealing top plate. The channel geometry can be designed in almost arbitrary configurations enabling the enhancement of the OPL while maintaining a compact device. The use of a small absorption cell is motivated by more than just its improved portability. A smaller sample gas volume allows for faster sampling, which is advantageous for real-time applications. Thus, the gas cell volume should be seen in relation to the OPL, making the volume-normalized path length $\delta$ = OPL/V an important figure of merit [16]. iHWGs proved already their potential in advanced sensing systems in the mid-infrared [17–19] and the ultraviolet [20].
Here, we present for the first time the implementation of iHWGs in a THz sensing system. We realize a compact spectroscopic platform with the potential of high versatility by the combination of robust iHWGs with the recently introduced flexible opto-electronic THz source enabling easy-handling at room temperature [21]. The direct electronic control of the adjustable spectral shape is implemented by the combination of optical sideband generation via electro-optic modulators [22] and difference frequency generation (also called photomixing) in a photoconductive THz antenna [23,24]. This enables a sideband (SB) tuning technique similar to other SB based spectroscopic methods [25–27]. In direct comparison, it surpasses the classical continuous wave single line sweep [28] in terms of speed, accuracy and straight-forward handling. The flexible sensing is demonstrated by measuring 9 different absorption lines of the important greenhouse gas nitrous oxide (100% N$_{2}$O). Thus, the combination of our frequency agile THz source enabling the SB method and the robust iHWGs providing inherent design flexibility, paves the way for THz sensing systems for real-world applications.
2. Methods
2.1 Experimental setup
The THz sensing platform is represented schematically in Fig. 1(a). The THz spectral components are defined completely in the NIR domain, starting with an intensity modulation applied to a 1555 nm, fiber-integrated, temperature stabilized NIR laser diode (LD1, Thorlabs SFL1550P, linewidth <200 kHz) in a fiber-coupled electro-optic intensity modulator (EOM, iXBlue MXAN-LN-20) driven by an RF signal generator (RFG, Agilent N5173B) with adjustable modulation frequency f$_\mathrm {m}$. The EOM is based on a Mach-Zehnder interferometer, thus, introducing a $\pi$/2 phase shift between both interferometer arms results in a dual sideband generation (SB$_{-}$=f$_\mathrm {LD1}$-f$_\mathrm {m}$, SB$_{+}$=f$_\mathrm {LD1}$+f$_\mathrm {m}$) with $\geq$15dB carrier suppression (f$_\mathrm {c}$=f$_\mathrm {LD1}$) as indicated in the inset of Fig. 1(a). For the generation of THz light, the modulated NIR radiation is combined with a second monochromatic NIR laser in a photomixing device, enabling a difference frequency mixing. For that purpose, a tunable NIR diode laser (LD2, ClarityPlus CLX-C, linewidth <500kHz) - referenced to an internal HCN gas cell [29] - is installed in the setup and all the NIR radiation is coupled in a 50:50 fiber combiner. The combined NIR light is boosted to 25 mW by an erbium doped fiber amplifier (EDFA, Thorlabs EDFA100P) and fed to a photoconductive THz antenna acting as mixer (PCA, Toptica EK000724). The PCA represents an efficient opto-electronic NIR to THz converter based on the optical beating of the NIR spectral lines at the PCA that induces a modulation of the photo current at THz frequencies, which drives the resonant antenna structure and, thus, emits THz radiation into free space. The offset between LD1 and LD2 together with the modulation frequency f$_\mathrm {m}$ defines uniquely the resulting THz frequencies ($\Delta$f=f$_\mathrm {LD2}$-f$_\mathrm {LD1}$, $\Delta$f-f$_\mathrm {m}$, $\Delta$f+f$_\mathrm {m}$). Two completely independent NIR lasers are used with a standard temperature stabilisation leading to a drift of $\Delta$f in the range of $\sim$50 MHz observed within a 10 min time window. The standard fiber-optic components and the electronic control of the synthetized THz spectra lead to a robust source providing flexible tuning and straightforward handling with a spectral resolution of <10 MHz [21]. The THz light is coupled via two off-axis parabolic mirrors (OPM) with a focal length of 100 mm to a substrate-integrated hollow waveguide (iHWG) fabricated with in- and outlet for gaseous species. A pressure gauge (Keller LEO 2) is used to monitor the pressure in the gas cell and its variation. During the data acquisition in a 1 min time window, a variation of $\sim$2 mbar is measured. A Golay cell (Tydex Ltd.) with a responsivity of 70 kV/W and a noise-equivalent power (NEP) of 120 pW/Hz$^\mathrm {1/2}$ acts as THz power detector and is placed directly at the output of the iHWG without the need of additional optics for coupling. For the sensing experiments the iHWGs are equipped with gallium arsenide (GaAs) or polyethylene (PE) windows to enable THz radiation propagation and to seal the gas sample volume. A sketch of an iHWG is depicted in Fig. 1(b) and optical pictures of two adapted iHWGs are presented in Fig. 1(c) and (d).
Fig. 1. (a) Sketch of the THz sensing platform consisting of a near-infrared dual sideband (DSB) generator realized by a fixed-wavelength diode laser (LD1) and an electro-optic intensity modulator (EOM) driven by an RF generator (RFG), a tunable diode laser (LD2), an erbium doped fiber amplifier (EDFA), a photodiode (PD) and a photoconductive antenna (PCA) acting as NIR to THz converter. The generated THz light is coupled via two off-axis parabolic mirrors (OPM) to a substrate-integrated hollow waveguide (iHWG) guiding the beam to a THz power detector. The inset shows schematically the DSB generation (SB$_{-}$, SB$_{+}$) with suppressed carrier (f$_\mathrm {c}$). (b) Schematic drawing of an iHWG. Example iHWGs equipped with gallium arsenide (c) and polyethylene (d) windows, respectively.
Five iHWGs with varying dimensions and optical path lengths are used in this work. They exhibit straight-line optical channels integrated into an aluminium (Al) substrate and cut-off frequencies calculated by [30]
Table 1. Optical channel and footprint dimensions (length x width x depth) of the iHWG devices used in this work and their calculated cut-off frequencies.a
2.2 Gas sensing techniques
Two different sensing techniques are compared in the following, namely the laser diode (LD) tuning and the sideband (SB) tuning. The LD tuning is a standard continuous wave (cw) single frequency sweep [28], realized by adjusting the wavelength of LD2 without any intensity modulation, while LD1 remains at fixed wavelength. The resulting single THz frequency is tuned through an absorption region while recording the transmitted intensity, which represents directly a measure of the line shape of the molecular transition under investigation. Each new frequency that is set at LD2 needs to be referenced to the internal HCN gas cell leading to measurement times of $\sim$20 min.
As an alternative, we introduce a SB tuning method comparable to other SB based spectroscopic techniques [25–27], which relies solely on the electronic tuning of the RF modulation frequency f$_\mathrm {m}$ driving the EOM. Figure 2(a) depicts the sweep of the generated sidebands (SB$_{-}$, SB$_{+}$) simultaneously along both slopes of an absorption line of a gas sample, while the center frequency (f$_\mathrm {c}$) remains fixed to the middle of that line. The alignment of f$_\mathrm {c}$ to the center of that line is guaranteed by the frequency referencing of LD2 to the internal available grid of HCN lines. The frequency drift of the free running LD1 is referenced to the nearest HCN line via optical beating of LD1 and LD2 on a fast photo diode (PD). During a measurement it remains below 50 MHz. For the actual gas measurement the offset is tuned to the desired THz frequency marking the center of the absorption line under investigation via further adjustment of the LD2 wavelength. In this way all frequencies involved are determined enabling the reconstruction of the actual THz frequency scale without the need of a spectrometer. The tunability of the LD2 wavelength enables the alignment of f$_\mathrm {c}$ to any absorption line in the accessible frequency window up to 3.5 THz limited by the frequency response of the used THz photoconductive antenna [24]. The radio-frequency control of the SB frequency offset is fast, convenient and provides an excellent precision (<10 Hz) leading to a measurement time of $\sim$1 min (or a data acquisition time of $\sim$0.6 s per frequency point), mainly determined by the used slow THz detector. This increase in measurement speed is crucial for potential applications such as in situ environmental monitoring or medical screening.
Fig. 2. (a) Schematic depiction of the SB tuning method. Adjustment of the RF modulation frequency f$_\mathrm {m}$ sweeps the sidebands (SB$_{-}$, SB$_{+}$) simultaneously along both slopes around the center of an absorption line. (b) Simulation of the SB tuning results in a direct measure of the right slope of a symmetric absorption line (purple dots). Reconstruction of the full line shape (grey dots) enables a Lorentzian fit (light blue line).
A simulation of a measurement with the SB tuning is shown in 2(b). The total transmission of all three THz frequencies (i.e. the suppressed carrier and the two sidebands) is recorded together at the THz detector, while each is experiencing an absorption strength depending on its position on the frequency scale in respect to the absorption line. This results essentially in a superposition of the right and the left half of that line (purple dots) with a negligible offset introduced by the suppressed carrier, which remains fixed to the center of the absorption line. Due to the symmetric nature of the Lorentzian line shape of a gas sample under ambient condition (where pressure broadening is the dominant broadening mechanism), the recorded data represents the direct measure of the right half of the line under investigation. Thus, the total transmission data can be used to reconstruct the actual line shape (grey dots). It should be noted that in the actual experiment the reconstruction also copies the noise of the measured data to the left half of the line and that the used RF equipment does not support frequencies below $\sim$300 MHz leading to a blind spot around the center of the line. From the full line shape the rotational transition properties (i.e. linewidth and absorption coefficient) are extracted via a Lorentzian fit (light blue line).
3. Results
3.1 iHWG characteristics
To examine the suitability of iHWGs for THz applications, 5 devices with varying dimensions and optical path lengths (overview in Table 1) are studied regarding their transmission properties. The LD tuning method is used for the investigation, since it provides absolute transmission values for each frequency. The transmitted signal through the iHWG is compared to the intensity recorded when the iHWG is removed (and the detector is moved to the position of the entry facet of the iHWG) acting as reference. Two of the waveguides are fabricated with an optical funnel at one end (i.e. a widening taper of the optical channel at the very end) to facilitate light coupling. To ensure equal incoupling efficiencies for same dimensions, those iHWGs were implemented in the setup with the optical funnel facing the detector.
Transmission data were obtained from 0.2 to 1.6 THz and are shown in Fig. 3(a). The observed dispersion follows the theory of rectangular metallic hollow waveguides with the cut-off frequencies calculated in Table 1. The discrepancy in the roll-off at low frequencies for different cross sections can also be explained by the f$_\mathrm {cut}$ values that are higher for the smaller waveguide dimensions. The frequencies used during the experiments are >3$\cdot$f$_\mathrm {cut}$ (for dimensions 2x2 mm) and >8$\cdot$f$_\mathrm {cut}$ (for dimensions 4x4 mm), respectively, considering the fundamental f$_\mathrm {cut}$. Furthermore, the measurements reveal losses while coupling to the iHWGs of $\sim$40$\%$ and $\sim$80$\%$, respectively, that remain similar for equal dimensions. It can also be seen that the optical funnels do not have a significant impact on the coupling from the waveguide to the detector.
Fig. 3. (a) Transmission through five different iHWGs with varying dimensions (summarized in Table 1) measured in the range of 0.2 to 1.6 THz. Fabry-Perot resonances measured in Al-90-GaAs (b) and Al-30-PE (c), respectively.
THz light traveling along the waveguide experiences the propagation loss $\mathcal {L}$ [dB/cm]. Pairs of the previous discussed transmission measurements (T$_\mathrm {i}$, T$_\mathrm {j}$) in iHWGs with different lengths (L$_\mathrm {i}$, L$_\mathrm {j}$) - but equal cross section - are used to calculate
An upper margin for the mean propagation loss for the whole frequency range of $\bar {\mathcal {L}}$ <0.14 dB/cm can be defined, its true nature is smaller than the error margins of the measurements. Thus, the propagation losses in the studied iHWGs are low enough to qualify them for proper wave guiding in the THz domain.
Al-90-4 and Al-30-4 have been chosen for further measurements due to their lower f$_\mathrm {cut}$ and smaller coupling losses compared to the 2x2 mm cross section iHWGs. They have been equipped with gallium arsenide (GaAs) and polyethylene (PE) windows (Al-90-4 $\rightarrow$ Al-90-GaAs and Al-30-4 $\rightarrow$ Al-30-PE) to enable THz radiation propagation and gas tightness at the same time. They provide a volume-normalized path length of $\delta$=6.2$\cdot$10$^{4}$ and require gas volumes of 1.44 mL for the 90 mm cell and 0.48 mL for the 30 mm cell. Remeasuring the transmission through those waveguides reveal Fabry-Pérot cavity resonances that can be seen in Fig. 3(b) and (c). Data fitting leads to a free spectral range (FSR) of 1.65 GHz for Al-90-GaAs and 4.76 GHz for Al-30-PE. Those values are in good agreement with the FSR calculated based on the physical length of the iHWGs. For Al-30-PE one needs to take into account that the 1 mm thick PE windows are part of the effective cavity due to the low Fresnel reflectivity coefficient of PE. Additionally, it can be seen that the GaAs windows induce a larger modulation depth and a lower overall transmission due to the higher reflectivity compared to the PE windows.
3.2 Comparison of the LD and SB tuning performance
For the comparison of both tuning techniques with respect to their suitability for gas sensing in iHWGs, both methods are studied to investigate the accuracy of the tuning mechanism itself and their corresponding signal fluctuations coming from the system instabilities. For that purpose, the total transmitted signal through an empty iHWG is recorded for every frequency. Figure 4 presents the measured data for three consecutive sweeps (S$_{1}$, S$_{2}$, S$_{3}$) of the LD wavelength (a) and the SBs (b) in the upper panels, while the lower panels compare all sweeps to the average intensity measured at every frequency, revealing a difference of up to one order of magnitude in the residual signals. The frequency accuracy of the LD tuning technique is mainly determined by the stability of the laser diodes, which exhibit linewidths of <500 kHz and are susceptible to temperature drifts, whereas the generation of SBs is conveniently controlled by the RFG with an accuracy on the Hz level. Overall, the photometric accuracy of the SB tuning mechanism exceeds that of the LD tuning mechanism by a multiple. Additionally, the SB tuning technique decreases the measurement times by a factor of 20 making the method robust against slow drifts of e.g. the temperature.
Fig. 4. In three consecutive sweeps (S$_{1}$, S$_{2}$, S$_{3}$) with (a) the standard LD tuning method and (b) the SB tuning technique, the total transmitted intensity through the empty Al-30-PE is recorded (upper panels). The residuals (lower panels) for every sweep show that the SB tuning outperforms the LD tuning in terms of photometric accuracy.
3.3 Gas sensing
We have chosen rotational transitions of the greenhouse gas nitrous oxide (N$_{2}$O) between 376.777 GHz (12.56792 cm$^{-1}$) and 702.983 GHz (23.44899 cm$^{-1}$) to demonstrate the sensing capability of the iHWGs in the THz regime. The transmission through Al-90-GaAs and Al-30-PE continuously flushed (gas flow of 0.1 L/min) with N$_{2}$O (N5.0 purity) is measured at various pressures between 150 and 300 mbar at an ambient temperature of 296 K.
In Fig. 5(a) a measurement of an individual N$_{2}$O absorption line in Al-90-GaAs using the SB tuning method with a frequency step size of 25 MHz is shown. The transmission through the iHWG filled with the gas sample is recorded and represents a direct measure of the right half of the N$_{2}$O line (blue dots) as described in the methods section. The reconstruction of the full lineshape (grey dots) enables the extraction of the linewidth and the absorption coefficient of the measured rotational transition via a Lorentzian fit (red line) revealing values of FWHM=1332.92$\pm$54.28 MHz and $\alpha$=14.0$\pm$1.33e$^{-3}$ cm$^\mathrm {-1}$. The deviation of the data from the Lorentzian fit (i.e. the fitting error) determines the error margins of the measurement. $\alpha$ is one order of magnitude smaller as the previously shown propagation losses, which is accessible due to the decreased error margins of the SB technique and the extraction from the whole line shape instead of measuring the amplitude of the absorption only.
Fig. 5. (a) Transmission measured (blue dots) at the N$_{2}$O line at 401.885 GHz (13.405428 cm$^{-1}$) with the SB tuning in Al-90-GaAs at a pressure of 210 mbar and a temperature of 296 K. Reconstruction of the full line shape (grey dots) enables the extraction of the linewidth and the absorption coefficient $\alpha$ via a Lorentzian fit (red line). (b) Measured pressure broadening of the absorption line with the LD tuning (green crosses) and the SB tuning (purple crosses) in Al-90-GaAs compared to the linewidth values (light blue line) obtained from the HITRAN database [3].
For the absorption line at 401.885 GHz (13.40543 cm$^{-1}$), Fig. 5(b) shows the well expected linewidth broadening with increasing N$_{2}$O pressure that follows the values obtained from the HITRAN database [3]. Using the SB tuning method, we obtained a pressure-induced broadening of 6.30$\pm$0.20 MHz/mbar, while the LD tuning method yields 5.48$\pm$1.11 MHz/mbar. Calculations based on the HITRAN database result in a value of 6.37 MHz/mbar, which is lying within the error margins of the experimental data for both methods. The measurements utilizing the SB tuning exhibit fitting errors $\sigma$ 4-5 times smaller compared to those of the LD tuning.
Figure 6 and the corresponding data in Table 2 demonstrate the flexibility of the presented sensing platform by targeting 9 randomly chosen N$_{2}$O lines. The extraction of the absorption coefficients on those lines reveal substantially decreased fitting errors for the measurements carried out with the SB tuning method, likewise to the linewidth results. Finally, it is noteworthy that the SB method provides unprecedented real-time control of the probing frequency, a feature further favouring this technique over a wavelength tuning based method. From these measurements a detection limit estimation for the presented demonstrator version of the sensing platform can be done. Considering the strongest N$_2$O absorption line measured at 23.448992 cm$^{-1}$ with a line intensity of 3.11e$^{-22}$ cm$^{-1}$/(molecule cm$^{-2}$) [3] and assuming a S/N ratio of 200 in view of the experiments carried out, it would be possible to detect 600 nmol of N$_2$O molecules present in the gas cell.
Fig. 6. Summary of absorption coefficients extracted from several measurements on 9 different N$_{2}$O lines in Al-90-GaAs (a) and Al-30-PE (b) with the LD tuning (green crosses) and the SB tuning (purple crosses) compared to the analogous HITRAN data [3] (light blue dots and lines). Corresponding data can be found in Table 2.
Table 2. Absorption coefficients extracted from measurements on 9 different N$_{2}$O lines in Al-30-PE and Al-90-GaAs (marked with $^{*}$) carried out with the two sensing methods.a
For the accurate quantification of unknown gaseous samples in real-world sensing applications, the determination of the absorption coefficient and the linewidth is essential. The SB technique proved to be useful for both tasks, significantly exceeding the results of the LD tuning method, making it a promising technique for potential sensing applications such as exhaled breath diagnostics [3,31].
4. Conclusion
In conclusion, we present for the first time the implementation of substrate-integrated hollow waveguides in the THz domain. They show low propagation losses suitable for guiding THz waves, which is demonstrated by measuring absorption coefficients of various rotational transitions of the greenhouse gas N$_{2}$O and by observing the collisional broadening of the linewidth of one of those lines, all in good agreement with the HITRAN data [3]. The combination with our electro-optic THz source enables a SB tuning method that outperforms the compared classical LD tuning technique in terms of measurement times (1 min vs. 20 min), accuracy (error margins in the range of 1-12% vs. 12-20%), robustness against slow drifts and handling convenience. The compactness of the iHWGs comes with the advantage of small sample volumes less than 2 mL and their volume-normalized optical path length of 6.2$\cdot$10$^{4}$, which compares very well to those of multipass gas cells [7,16]. iHWGs show the potential to play a key role on the way to practical THz sensing applications providing mechanical robustness and small footprints at low production cost. The low propagation losses open opportunities for extended optical path lengths with meandring structures instead of straight light channels or even more advanced concepts [32]. Their integration possibilities might be fully exploited implementing transmitter and receiver directly into the substrate progressing the development of portable sensing devices [20,33].
The inherent design flexibility of iHWGs with their robust- and compactness and their simple production, together with the electro-optic THz source already proven to provide straightforward synthesis of precise custom THz spectra up to 3.5 THz [21] and the fast SB technique without the necessity of any laser cavity tuning or permanent frequency referencing assemble a powerful platform that represents another step on the way to a highly versatile THz sensing platform adaptable to a wide range of real-world application scenarios, especially in areas relying on small gas volume sampling.
Funding
Austrian Science Fund (DiPQCL P30709-N27, DK CoQuS W1210).
Acknowledgments
The authors acknowledge Thomas Müller (TU Wien) for providing one of the diode lasers and Michael Feiginov (TU Wien) for helping with room temperature THz detection. BM and MH acknowledge support by the DFG Graduiertenkolleg (GRK) 2203 (PULMOSENS) and by the Ministerium für Wissenschaft, Forschung und Kunst (MWK) in Baden-Württemberg, Germany within the project ’IRENE’.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are available from the authors upon reasonable request.
References
1. N. Rothbart, O. Holz, R. Koczulla, K. Schmalz, and H.-W. Hübers, “Analysis of Human Breath by Millimeter-Wave/Terahertz Spectroscopy,” Sensors 19(12), 2719 (2019). [CrossRef]
2. A. Cuisset, F. Hindle, G. Mouret, R. Bocquet, J. Bruckhuisen, J. Decker, A. Pienkina, C. Bray, E. Fertein, and V. Boudon, “Terahertz Rotational Spectroscopy of Greenhouse Gases Using Long Interaction Path-Lengths,” Appl. Sci. 11(3), 1229 (2021). [CrossRef]
3. I. E. Gordon, L. S. Rothman, R. J. Hargreaves, et al., “The HITRAN2020 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transf. 277, 107949 (2022). [CrossRef]
4. J. Hodgkinson and R. P. Tatam, “Optical gas sensing: a review,” Meas. Sci. Technol. 24(1), 012004 (2013). [CrossRef]
5. J. D. Cruzan, M. R. Viant, M. G. Brown, D. D. Lucas, K. Liu, and R. J. Saykally, “Terahertz laser vibration-rotation-tunneling spectrum of the water pentamer-d10.: Constraints on the bifurcation tunneling dynamics,” Chem. Phys. Lett. 292(4-6), 667–676 (1998). [CrossRef]
6. C. Hepp, S. Luttjohann, A. Roggenbuck, A. Deninger, S. Nellen, T. Gobel, M. Jorger, and R. Harig, “A cw-terahertz gas analysis system with ppm detection limits,” 41st International Conference on Infrared, Millimeter, and Terahertz Waves pp. 1-2 (2016).
7. N. Rothbart, K. Schmalz, and H.-W. Hübers, “A Compact Circular Multipass Cell for Millimeter-Wave/Terahertz Gas Spectroscopy,” IEEE Trans. Terahertz Sci. Technol. 10(1), 9–14 (2020). [CrossRef]
8. F. Hindle, R. Bocquet, A. Pienkina, A. Cuisset, and G. Mouret, “Terahertz gas phase spectroscopy using a high-finesse Fabry-Pérot cavity,” Optica 6(12), 1449–1454 (2019). [CrossRef]
9. M. S. Islam, C. M. B. Cordeiro, M. A. R. Franco, J. Sultana, A. L. S. Cruz, and D. Abbott, “Terahertz optical fibers [Invited],” Opt. Express 28(11), 16089–16117 (2020). [CrossRef]
10. S. Atakaramians, S. A. V T. M. Monro, and D. Abbott, “Terahertz dielectric waveguides,” Adv. Opt. Photonics 5(2), 169–215 (2013). [CrossRef]
11. B. You, T.-A. Liu, J.-L. Peng, C.-L. Pan, and J.-Y. Lu, “A terahertz plastic wire based evanescent field sensor for high sensitivity liquid detection,” Opt. Express 17(23), 20675–20683 (2009). [CrossRef]
12. M. S. Islam, J. Sultana, K. Ahmed, M. R. Islam, A. Dinovitser, B. W.-H. Ng, and D. Abbott, “A Novel Approach for Spectroscopic Chemical Identification Using Photonic Crystal Fiber in the Terahertz Regime,” IEEE Sens. J. 18(2), 575–582 (2018). [CrossRef]
13. J. Sultana, M. S. Islam, K. Ahmed, A. Dinovitser, B. W.-H. Ng, and D. Abbott, “Terahertz detection of alcohol using a photonic crystal fiber sensor,” Appl. Opt. 57(10), 2426–2433 (2018). [CrossRef]
14. A. L. S. Cruz, C. M. B. Cordeiro, and M. A. R. Franco, “3D Printed Hollow-Core Terahertz Fibers,” Fibers 6(3), 43 (2018). [CrossRef]
15. A. Wilk, J. C. Carter, M. Chrisp, A. M. Manuel, P. Mirkarimi, J. B. Alameda, and B. Mizaikoff, “Substrate-Integrated Hollow Waveguides: A New Level of Integration in Mid-Infrared Gas Sensing,” Anal. Chem. 85(23), 11205–11210 (2013). [CrossRef]
16. M. Graf, L. Emmenegger, and B. Tuzson, “Compact, circular, and optically stable multipass cell for mobile laser absorption spectroscopy,” Opt. Lett. 43(11), 2434–2437 (2018). [CrossRef]
17. J. F. da Silveira Petruci, P. R. Fortes, V. Kokoric, A. Wilk, I. M. Raimundo Jr., A. A. Cardoso, and B. Mizaikoff, “Monitoring of hydrogen sulfide via substrate-integrated hollow waveguide mid-infrared sensors in real-time,” Analyst 139(1), 198–203 (2014). [CrossRef]
18. I. J. G. da Silva, E. Tütüncü, M. Nägele, P. Fuchs, M. Fischer, I. M. Raimundo, and B. Mizaikoff, “Sensing hydrocarbons with interband cascade lasers and substrate-integrated hollow waveguides,” Analyst 141(14), 4432–4437 (2016). [CrossRef]
19. J. Glöckler, C. Jaeschke, Y. Kocaöz, V. Kokoric, E. Tütüncü, J. Mitrovics, and B. Mizaikoff, “iHWG-MOX: A Hybrid Breath Analysis System via the Combination of Substrate-Integrated Hollow Waveguide Infrared Spectroscopy with Metal Oxide Gas Sensors,” ACS Sens. 5(4), 1033–1039 (2020). [CrossRef]
20. D. N. Barreto, W. R. Silva, B. Mizaikoff, and J. F. da Silveira Petruci, “Monitoring Ozone Using Portable Substrate-Integrated Hollow Waveguide-Based Absorbance Sensors in the Ultraviolet Range,” ACS Meas. Sci. Au 2(1), 39–45 (2022). [CrossRef]
21. D. Theiner, B. Limbacher, M. Jaidl, K. Unterrainer, and J. Darmo, “Flexible terahertz opto-electronic frequency comb light source tunable over 3.5 THz,” Opt. Lett. 46(22), 5715 (2021). [CrossRef]
22. V. Torres-Company and A. M. Weiner, “Optical frequency comb technology for ultra-broadband radio-frequency photonics,” Laser Photonics Rev. 8(3), 368–393 (2014). [CrossRef]
23. E. R. Brown, K. A. McIntosh, K. B. Nichols, and C. L. Dennis, “Photomixing up to 3.8 THz in low-temperature-grown GaAs,” Appl. Phys. Lett. 66(3), 285–287 (1995). [CrossRef]
24. T. Göbel, D. Stanze, B. Globisch, R. J. B. Dietz, H. Roehle, and M. Schell, “Telecom technology based continuous wave terahertz photomixing system with 105 decibel signal-to-noise ratio and 3.5 terahertz bandwidth,” Opt. Lett. 38(20), 4197–4199 (2013). [CrossRef]
25. G. C. Bjorklund, “Frequency-modulation spectroscopy: a new method for measuring weak absorptions and dispersions,” Opt. Lett. 5(1), 15–17 (1980). [CrossRef]
26. P. Verhoeve, E. Zwart, M. Versluis, M. Drabbels, J. J. ter Meulen, W. L. Meerts, A. Dymanus, and D. B. McLay, “A far infrared laser sideband spectrometer in the frequency region 550-2700 GHz,” Rev. Sci. Instrum. 61(6), 1612–1625 (1990). [CrossRef]
27. R. Eichholz, H. Richter, M. Wienold, L. Schrottke, R. Hey, H. T. Grahn, and H.-W. Hübers, “Frequency modulation spectroscopy with a THz quantum-cascade laser,” Opt. Express 21(26), 32199–32206 (2013). [CrossRef]
28. A. S. Pine, R. D. Suenram, E. R. Brown, and K. A. McIntosh, “A Terahertz Photomixing Spectrometer: Application to SO2Self Broadening,” J. Mol. Spectrosc. 175(1), 37–47 (1996). [CrossRef]
29. S. L. Gilbert, W. C. Swann, and C.-M. Wang, “Hydrogen Cyanide H13C14N Absorption Reference for 1530 nm to 1560 nm Wavelength Calibration SRM 2519,” NIST pp. 37–38 (1998).
30. J. D. Jackson, Classical Electrodynamics (John Wiley & Sons, 1998).
31. I. A. Ratiu, T. Ligor, V. Bocos-Bintintan, C. A. Mayhew, and B. Buszewski, “Volatile Organic Compounds in Exhaled Breath as Fingerprints of Lung Cancer, Asthma and COPD,” J. Clin. Med. 10(1), 32 (2020). [CrossRef]
32. D. N. Barreto, V. Kokoric, J. F. da Silveira Petruci, and B. Mizaikoff, “From Light Pipes to Substrate-Integrated Hollow Waveguides for Gas Sensing: A Review,” ACS Meas. Sci. Au 1(3), 97–109 (2021). [CrossRef]
33. V. Kokoric, P. A. Wissel, A. Wilk, and B. Mizaikoff, “muciPRECON: multichannel preconcentrators for portable mid-infrared hydrocarbon gas sensors,” Anal. Methods 8(36), 6645–6650 (2016). [CrossRef]
