Open Access
Add to BibSonomy
Abstract
An off-beam quartz-enhanced photoacoustic spectroscopy (QEPAS) sensor was designed for ethylene detection using a distributed-feedback quantum cascade laser (QCL) operating in the mid-infrared around 11 μm. The acoustic microresonator configuration was experimentally optimized using an original open-cell photoacoustic setup with a MEMS microphone. Correction factors based on theoretical acoustic models were introduced in order to accurately describe the response of millimeter-sized acoustic resonators. The optimized QEPAS sensor exhibited a limit of detection of 60 ppb for 60 s integration, giving a NNEA of 4.8 × 10−8 W·cm−1·Hz-0.5.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Quartz-enhanced photoacoustic spectroscopy (QEPAS) has been largely used in numerous applications since its invention by A. Kosterev in 2002 [1]. This technique offers many advantages compared to other spectroscopic methods in terms of simplicity and sensitivity which participated to its success and development.
QEPAS is based on the photoacoustic effect using tunable laser sources. The light of the source can be absorbed by a specific gaseous specie when an absorption line corresponds to the laser emitted wavelength. If the light source is modulated in intensity, the absorption is also modulated and gives rise to a localized sound wave interacting with a quartz tuning fork (QTF), leading to a piezoelectric current.
In order to enhance the pressure wave amplitude, small tubes called micro-resonators (mR) are added. In the on-beam configuration the resonator tubes are placed on both sides of the QTF leaving a gap between the QTF prongs whereas in the off-beam configuration the resonator tube is located on one side of the QTF, alongside the prongs [2]. On-beam QEPAS exhibits a better signal enhancement factor, around 30 [3,4], but it requires a stringent optical alignment to avoid photo-thermal effects due to straight illumination of the QTF. Despite a smaller signal enhancement, the off-beam configuration reduces optical constraints. This configuration was theoretically studied in [5] and successfully used for molecular spectroscopy [6–9]. This is the one we use in the present study.
Ethylene (C2H4) is the organic product produced in the highest quantities by the industry and its production is increasing. It is used to produce many polymers (e.g. polyethylene, polystyrene), detergents, solvents and antifreeze. Ethylene is also a plant hormone involved in complex bio-regulation processes [10,11], which stimulate the maturation of climacteric fruits. Melons are widely produced in the southern regions of France. Maturity assessment, through continuous ethylene monitoring, can be a quantitative indicator for convenient harvesting and storage. The ripening of fruits can also be adjusted by controlling the ethylene concentration in an air-controlled chamber, where a concentration of a few ppmv is enough to activate the maturation processes. Electro-catalytic sensors are sometimes used for ethylene monitoring but they do not provide the necessary selectivity and accuracy.
QEPAS has already been used for ethylene detection. A 1.62 µm distributed feedback (DFB) laser, with an average optical power of 15 mW, in the on-beam configuration was first used [12]. A noise equivalent signal of 4 ppmv was achieved for 0.7 s integration time on an absorption peak at 6177.14 cm−1. Later a quantum cascade laser was employed to detect ethylene at 10.5µm [13], here again in the on-beam configuration. A detection limit of 50 ppbv in 70s was achieved using second harmonic measurements, giving a normalized noise equivalent absorption (NNEA) of 1.78 × 10−7 W·cm−1·Hz-0.5. The influence of CO2 on ethylene sensing at low pressure was then evaluated using the same laser [14].
In this paper we present a QEPAS ethylene sensor using a 915 cm−1 (10.9µm) InAs/AlSb DFB quantum cascade laser fabricated in our laboratory [15]. We show the acoustic characterization and optimization of the microresonators used in the off-beam configuration, using a commercial MEMS microphone. We performed ethylene measurements using calibrated mixtures and evaluated the detection limit of the sensor.
2. Laser specifications and wavelength selection
The QCL used in this study was based on the InAs/AlSb materials and grown by Molecular Beam Epitaxy (MBE) in a RIBER 412 solid-source MBE system. The laser structure is reported in [15]. The grown wafer was processed into 7-µm-wide ridge lasers. A linear grating was patterned on the top of the ridges using electron beam lithography and inductively coupled plasma etching. The ridge surface was then metallized with gold thus forming a metal DFB grating in the laser waveguide. The periodicity of the 1st order grating was selected to be 1.63 µm, giving an emission wavelength close to the maximum of the QCL gain spectrum. The fabricated devices were mounted epi-side down on copper heatsinks. The lasers operated in the continuous wave regime up to room temperature with a threshold current density of 1.3 kA/cm2. The voltage-current and light-current characteristics of the laser at different temperatures are shown on Fig. 1.
Fig. 1 Voltage-current and light-current characteristics (a) and emission spectra (b) of a 3.6-mm-long QCL at different temperatures. The ethylene absorption spectrum, on top of Fig. 1(b), shows many features in the range covered by the QCL.
For spectral characterization, a Fourier transform infrared spectrometer combined with a pyroelectric detector was used. The laser emission could be tuned by varying both the current and the temperature, with respective tuning rates of −0.9 cm−1/A and −0.09 cm−1/K. For our purpose, the laser temperature was chosen to be around 240K for two reasons: the emission wavelength is close to a strong absorption peak and the optical power is up to 3 mW. The measurements were performed using a homemade Peltier temperature controlled module capable to cool the laser at temperatures down to 240K without using liquid nitrogen, thus making the system more suitable for field deployment. In total, with a temperature span of 40K, a spectral range of about 3.5 cm−1 can be covered, addressing various ethylene absorption lines.
As shown on Fig. 2, the composite ethylene absorption spectrum results from multiple transitions with linestrengths ranging from 10−21 to 10−19 cm−1/(mol·cm2). Those bands add up to form a double peak with maxima located at 915.61 and 915.25 cm−1. At a concentration of 1 ppmv, it exhibits an absorption of 7.94 × 10−6 cm−1 (a factor 1/5 compared to the strongest peak at 951 cm−1) and a half width at half maximum of about 0.28 cm−1. In terms of common interfering species, the closest H2O absorption is located far from the laser setpoint (913.97 cm−1). For a usual atmospheric composition (1% H2O, 450 ppm CO2, in volume), absorption due to CO2 was 1 order of magnitude lower than that of a 100 ppbv C2H4 target sample.
Fig. 2 Selected absorption peaks (green) obtained from the Hitran database [17] around 915.61 cm−1, for 1ppmv of C2H4 at standard conditions of pressure and temperature, and the corresponding absorption lines (purple).
Wavelength Modulation Spectroscopy (WMS) is one of the most sensitive techniques in spectroscopy. The laser is driven with a slow ramp to tune the laser central wavelength, which is additionally modulated with a high frequency sinusoidal component for the wavelength modulation. In QEPAS, the high frequency modulation corresponds to an harmonic of the QTF fundamental frequency f0. In our experiment, the laser was driven either in the 1f (f = f0) or 2f (f = f0/2) mode, for which two different temperature and current setpoints have been selected. From a theoretical point of view, as described by Arndt [16], the wavelength modulation at the nth harmonic gives a signal proportional to the nth derivative of the absorption. For the selected ethylene absorption line, 1f and 2f maxima can be obtained from the 1st and the 2nd derivative. Arndt’s theory stands for small modulation amplitude, and is to be refined in the case of large amplitude modulation.
3. Experimental setup of the sensor
A schematic of the QEPAS sensor for ethylene detection is presented in Fig. 3. As for a conventional QEPAS system, the photoacoustic generation is obtained using a DFB laser, and the detection is provided by a QEPAS spectrophone enclosed in a gas cell. The spectrophone, consisting of the microresonator (mR) and the QTF, is based on an off-beam geometry, adapted to the QCL optical requirements.
Fig. 3 Schematic of the QEPAS setup (top view) for ethylene detection. The excitation source is a Quantum Cascade Laser (QCL). The acoustic spectrophone is made of a microresonator (mR) and a Quartz Tuning Fork (QTF).
The QCL emits an elliptic beam with a high divergence estimated to be 80x120° in the lateral and vertical directions, respectively. There are two critical points in the QEPAS optical design: efficient light collection, and preventing the direct illumination of the mR walls. To collect the maximum optical power, the collimation lens must be chosen with a high numerical aperture. However, aberrations at high angles can result in photothermal noise due to the illumination of the mR walls. The issue can be overcome by using a two-lens optical system, forming an intermediate image between the lenses and removing divergent rays with a pinhole [13]. Another approach employing an hollow-core fiber was successfully implemented [18]. In our case, for convenience of optical procedures, a single aspheric lens (L1) (Thorlabs C036TME-F, 4.0mm focal length) was used for laser light collection and focusing. The laser light is focused in the center of the main tube of the mR. At the entrance of the mR, the radius of the beam size is of 0.45mm. A 2-mm-diameter pinhole (P) is positioned at the entrance of the gas cell for coarse beam cleaning. A pyroelectric detector (Infratec LIE-332f-66) is used for optical alignments. In the off-beam configuration, the mR offers multiple benefits to the system: (1) it increases the acoustic pressure (as in on-beam QEPAS), (2) it can be adapted to the beam width, (3) it protects the QTF from the stray light by separating light absorption and sound wave detection. After photoacoustic generation and amplification, the acoustic wave is converted into an electric signal by the QTF. The QTF is a standard electrical component, with a resonant frequency of 32.7 kHz and a high quality factor (Q ≈8000 in air). It acts as an acoustic pass-band filter and is responsible for the very high signal to noise ratio obtained with a such spectrophone. The QEPAS response can be enhanced with an efficient coupling between the mR and the QTF, depending on the matching of the resonant frequencies and the relative position of the elements. The design of the mR and the position of the QTF are discussed below. The electrical signal is processed through a transimpedance amplifier and then a lock-in amplifier.
The gas cell has a volume of 45 mm3 and is equipped with two 20° tilted ZnSe windows, and two outlets for gas circulation. A gas mixing system (Alytech GasMix Aiolos II) was used to provide accurate dilution of ethylene in nitrogen, operating at atmospheric pressure and at a constant flow rate of 2L/min.
4. Off-beam spectrophone optimization
In this section, the acoustic behavior of the microresonator is studied and adjusted to meet the requirements of the QCL. Then, the QTF position, relative to the mR, is optimized. For these characterizations, another infrared laser operating around 2.62μm is used to target a H2O absorption peak. The photoacoustic generation is realized in a simple open-cell configuration (no gas chamber).
4.1. Acoustic characterization
The QEPAS spectrophone can be acoustically enhanced by the addition of tubes with millimetric dimensions, called microresonators [3,4]. They can be modelized as unidirectional resonator, the excited resonance is longitudinal, compatible with the symmetry of the Gaussian laser beam. For our off-beam setup, the microresonator is made of two tubes arranged in a T-shape (Fig. 4). The laser beam is passed through the main tube. The acoustic generation in this tube provokes the creation of a standing wave. The opening, located at the pressure antinode, i.e. at the center of the tube (x = LmR/2), serves as a secondary acoustic source for the excitation of the QTF.
Fig. 4 3D Section view of the T-shape mR composed of the main tube (red) of length LmR and radius RmR and the opening (blue) of radius r0 and length t0.
As the first assumption, the resonant frequencies for the longitudinal modes of the tube are given by [19]:
fn is the resonant frequency of the nth longitudinal mode, c the speed of sound and Leff the effective length. Leff can be obtained from LmR by applying an end correction for open tubes [20]:CmR is the correction factor for the main tube.However, the resonant frequency is impacted by any geometrical change in the mR, for instance here the presence of an opening. Theoretical models using acoustic impedance based calculations have been presented for the original off-beam mR and the more refined T-shape model [5,9]. For a given resonant frequency f0, the effective length can be calculated:
S and ss are the cross section of respectively the main tube and the opening. teff is the effective length of the opening considering open-end corrections:with t0 physical length, Ct0 correction factor and r0 radius of the opening (see Fig. 4).Generally, the correction factors CmR and Ct0 are taken as 1,2 [19,20]. Here, we suggest to adjust these factors to the experimental results and thereby, obtain a semi-theoretical model with better accuracy.
In the past experiments, the acoustic response of on-beam mR was obtained from the deduction of the measured QEPAS signal with or without the mR [4,21]. However, regarding the sensitivity to the QTF position and the narrow QTF response compared to the one of the mR, this method seems to be of a relatively low accuracy. To overcome this issue, we have developed an experimental acoustic test bench for mR spectral response characterization. As shown on Fig. 5, the setup is composed of a GaSb quantum well laser diode, a mR and a MEMS microphone. This diode laser was intentionally chosen to target a water vapor absorption peak at 3816.09 cm−1(2.62μm), with a linestrength of 2.31 × 10−19 cm−1/(mol·cm2). The optical power of the laser was about 1mW. The humidity of the ambient air was used as a target gas for photoacoustic generation. The measurement time was short compared to variations of humidity. Hence, no gas cell was required to enclose the mR, avoiding acoustic modes arising from the resonances of the gas chamber. Most photoacoustic experiments employ standard audio microphone in the kHz range. The selected MEMS microphone (Knowles SPU0410LR5H-QB) was able to operate in the ultrasonic domain, suitable for the characterization of the QEPAS mR around 32.7kHz. A frequency sweep of the laser modulation was used to scan the acoustic range of interest. Demodulation by a lock-in amplifier was necessary to remove the incoming noise of the broadband microphone. The microphone was placed 0.75mm away from the mR opening (Fig. 5(b)), distance where the acoustic behavior of the mR was unaffected by the microphone volume.
Fig. 5 (a) Photography of the acoustic test bench with, from left to right, the 2.62μm laser enclosed in a temperature regulated module (1), the focusing lens (2), the mounted microresonator (3) facing the MEMS microphone (4) placed on a xyz translation stage, and a tilted photodiode (5). (b) Close-up picture of the mR and the microphone.
9 microresonators with various lengths and diameters were selected (Table 1), and fabricated from aluminium blocks by milling and drilling. Different diameters would allow to adjust the mR for the future use of lasers with various optical waist size. The parameters r0 and t0 are set to fixed values, 0.25 and 0.5mm respectively, for the all 9 mRs. The length is adjusted to match the mR resonance with the one of the QTF. The speed of sound is calculated for ambient air with standard conditions of temperature, humidity and gas density [22].
Table 1. Fabricated microresonators and their dimensions as defined Fig. 4.
The quality factors of the different mRs are in the range of 10, and were difficult to extract from our measurements because the microphone response is not perfectly flat. Calibration methods of the microphone are currently under discussion. The comparison of experimental and theoretical results for the different mRs is presented Fig. 6, showing the peak frequency response for the different mRs tested.
Fig. 6 Acoustic resonance for resonators from Table 1. The correction factors are adjusted to fit the experimental results. f0 is the QTF resonant frequency, i.e the target frequency for the mR.
The original expression of Eq. (3) with CmR = Ct0 = 1.2 is clearly off the experimental results by approximately 1400Hz. The experimental data were numerically fitted by the Eq. (3) setting CmR and Ct0 as variable parameters, giving as a result CmR = 1.5 and Ct0 = 1.9. The corrected theoretical model exhibits a good fit of the experimental data. For this model, the mean error over the 9 mRs tested, is of 200Hz, that is about a factor 7 error reduction compared to the original model. This corrected model will be a very convenient tool for dimensioning the microresonator in various environmental conditions and for any target frequency. r0 and RmR will be chosen, respectively, for optimum coupling with the QTF and depending on the laser beam width. Then, LmR will be adjusted to match with the QTF resonant frequency. In our case, RmR was chosen to be 0.5mm and the deduced length LmR was 5.2mm (with same t0 = 0.5mm and r0 = 0.25mm).
4.2. Optimization of the QTF position
After, the experimental dimensioning of the mR was completed, we studied the influence of the QTF position. Using the same setup and replacing the microphone by a QTF gave us a suitable setup to optimize the QTF position relatively to the mR. For on-beam QEPAS, it has been described theoretically and experimentally verified [23]. However, the behavior in an off-beam configuration can differ from that of the on-beam setup. The acoustic source being the opening of the mR instead of the direct pressure wave associated with the laser beam, the pressure around the QTF prongs has a different profile in space. The QEPAS signal is thus recorded while moving the QTF in the 3 space directions (Fig. 7).
Fig. 7 Amplitude of the QEPAS signal as a function of the relative position between the mR and the QTF, in the x (a), z (b) and y (c) directions. Illustration of the referential (d) with the mR located at the origin (x = y = z = 0) corresponding to the black cross.
In the x direction, the signal has a maximum at x equal to zero. The fork has a mirror symmetry relative to the yz plane, thus the signal is optimal for the condition of symmetry. The signal decreases to almost zero when one of the prong is placed in front of the mR opening, followed by an increase due to the excitation of the external part of the prong. In the z direction, the optimum is located 1 mm below the top of the QTF. It is a little below the ideal z position for the on-beam QEPAS (0.7mm) [23]. The diameter of the orifice being 0.5 mm, the acoustic energy is more widespread than in on-beam, thus displacing the optimum z position by 0.3mm. Finally, the y direction being the most crucial, the quality factor was determined as well as the QEPAS signal. The Q factor decreases from 8000 to 3000 at 20 μm, due to viscous damping effects in the vicinity of the wall of the mR. The QEPAS signal exhibits a maximum around 100 μm. In conclusion, the ideal position of the QTF, relative to the mR, was determined and corresponds to an optimum acoustic coupling between the QTF and the mR.
5. C2H4 sensors results
The QEPAS sensor was then employed for spectroscopic purposes and ethylene sensing. The laser temperature was continuously adjusted from −30 to 10°C while recording the 2f QEPAS signal. The photoacoustic generation is proportional to the optical power. To obtain a power independent spectrum, the QEPAS signal was normalized by the optical power. The normalized signal is compared to the 2nd derivative of the ethylene spectrum (Fig. 8), exhibiting a very nice concordance. This spectroscopic acquisition provided us meaningful information for: (1) the verification of the QCL behavior with temperature, (2) an accurate calibration of the tunability with temperature (3) the selection of the laser setpoint for optimum detection.
Fig. 8 The 2nd derivative of ethylene absorption (a) is compared to the normalized 2f QEPAS signal (b). The laser current was of 340mA and the modulation amplitude of 0.04cm−1. Ethylene concentration was set to 5%.
A good compromise of laser power and absorption strength was found at −26°C for the 2f mode. The intensity of the photoacoustic generation strongly depends on the modulation amplitude. Following Arndt’s theory, the m modulation index is defined as the ratio of the laser wavelength excursion (due to the wavelength modulation) over the half width at half maximum of the absorption peak [16]. The maximum of the signal is reached for m equal to 2 and 2.2 for, respectively, the 1f and 2f mode. This model proved to be valid within the frame of small modulation amplitudes and describes only the behavior for a single Lorentzian peak. When multiple peaks are in the same vicinity, they will each have a contribution to the total composite photoacoustic signal. For an accurate theoretical description of the signal at high modulation, Arndt’s signal might be modified by: (1) considering the contribution of each absorption peak (2) adding the amplitude modulation due to the current dependence of the power [24].
Practically, in order to optimize the photoacoustic signal, the modulation amplitude must be studied simultaneously with the laser central frequency. At small modulation amplitudes, two peaks at 915.25 and 915.61cm−1 can be observed (Fig. 9(a)), corresponding to the ethylene absorption profile. As the modulation amplitude is increased, the two features merge together to form a single broad peak with a magnified amplitude. Thus, under this condition, the photoacoustic signal is not proportional to the n-th derivative. For gas sensing purposes it seems judicious to work with high modulation amplitudes in order to maximize the QEPAS signal. For the 2f mode, the optimum conditions are represented by the grey dotted lines on Fig. 9(a). It seems unappropriated to deduce a modulation index for a such complex absorption. It can be noted that the modulation amplitude of 0.44cm−1 is greater than the FWHM of the whole peak, that is 0.28cm−1, overlapping the two original ethylene peaks.
Fig. 9 QEPAS signal measured as a function of the wavenumber, for different modulation amplitudes, for the 2f (a) and 1f mode (b). 2nd and 1st derivative of ethylene absorption are represented in (c) and (d): They give the small amplitude modulation (0.04 cm−1) shapes of the 1f and 2f QEPAS signals. The grey dotted lines shows the optimum working conditions for both modulation schemes.
The study presented for the 2f mode was replicated for the 1f mode using the right side of the absorption peak at 915.61cm−1, as shown Fig. 9(c). The measured optical power was of 2.8 mW. Due to the broad feature of the ethylene spectra, the whole 1f signal could not be covered through a current sweep. Therefore, only the maximum of the 1f signal was recorded and not the amplitude. The linearity was then verified, proving that the 1f maximum is a meaningful quantity. From a theoretical point of view, the photoacoustic signal is stronger in 1f mode. This can be seen experimentally: there is a factor 2 between the 1f and 2f signal for optimum conditions. Nonetheless, it is well known that a signal background, specifically originating from intensity modulation at solid-gas interfaces (cell windows, walls, …) is present in 1f mode and almost negligible in the 2f mode [19]. With a cell filled with pure N2, the background was estimated to be equivalent in 1f and 2f. From this observation, it was judged relevant to compare the performance of the sensor in both modes.
The linearity of the sensor was evaluated by monitoring the 1f QEPAS signal for different ethylene concentrations. The cell was successively filled with a calibrated gas concentration and flushed with N2 between each steps to check the recovery of the zero signal. The results are presented Fig. 10(a), showing very good linearity.
Fig. 10 (a) Response of the sensor versus ethylene concentration. The integration time is set to 0.1s, giving an absolute error of 1.5 ppm. (b) Allan-Werle deviation calculated from a 30 minutes acquisition for the 1f and 2f mode. The C2H4 concentration was 200ppm. The sensor was stable for 60s.
200 ppm of ethylene mixed with pure nitrogen were introduced in the gas cell. The signal was recorded for 30 minutes. For a given integration time, we calculated the Allan-Werle deviation (Fig. 10). The deviation exhibits a similar behavior for both 1f and 2f modes, with a -t-1/2 slope indicating dominant white noise. Long-term drift appears after 60s. The limit of detection is of 60 ppb in the 1f mode, giving a NNEA of 4.8 × 10−8 W·cm−1·Hz-0.5.
A future work will focus on the improvement of the long term stability of the sensor. In particular, the variability of the QTF environment (temperature, humidity, gas density) and the laser behavior will be studied on its impact on the overall system stability.
6. Conclusion
In this work, we have presented a QEPAS sensor for ethylene detection using a DFB QCL and an off-beam QEPAS spectrophone. The fabricated QCL was appropriate for gas sensing in terms of spectral behavior, optical power and operating temperature.
In order to study the acoustic behavior of the mR, we have developed an acoustic test bench based on water vapor photoacoustic generation in an open cell configuration and a MEMS microphone. Reshaping the original theoretical model for off-beam QEPAS by adjusting open-end correcting factors, we obtained a semi-empirical model for the resonant frequency as a function of the mR dimensions. The position of the QTF was studied as well. The optima were relatively similar to the on-beam configuration with minor variations inherent to the specific acoustic profile of the off-beam setup.
Based on those useful results, a microresonator was designed to meet the requirements of the QCL. The 2f QEPAS signal demonstrated a very good correspondence with the ethylene absorption spectrum. In terms of performances, both 1f and 2f modulation schemes were explored. In our specific conditions, the signal-to-noise ratio in 1f mode was larger by an order of magnitude than the one for the 2f mode. The sensor attained a limit of detection of 60ppb in 60s, leading to a NNEA of 4.8 × 10−8 W·cm−1·Hz-0.5.
Future works on the sensor will focus on the improvement of the long-term stability, the implementation of high power QCLs and integration.
Funding
Agence Nationale de la Recherche (MULTIPAS project [ANR-16-CE04-0012], French “Investment for the Future” program [EquipEx EXTRA, ANR-11-EQPX-0016]).
References
1. A. A. Kosterev, Y. A. Bakhirkin, R. F. Curl, and F. K. Tittel, “Quartz-enhanced photoacoustic spectroscopy,” Opt. Lett. 27(21), 1902–1904 (2002). [CrossRef] [PubMed]
2. K. Liu, X. Guo, H. Yi, W. Chen, W. Zhang, and X. Gao, “Off-beam quartz-enhanced photoacoustic spectroscopy,” Opt. Lett. 34(10), 1594–1596 (2009). [CrossRef] [PubMed]
3. L. Dong, A. A. Kosterev, D. Thomazy, and F. K. Tittel, “QEPAS spectrophones: Design, optimization, and performance,” Appl. Phys. B Lasers Opt. 100(3), 627–635 (2010). [CrossRef]
4. H. Yi, W. Chen, X. Guo, S. Sun, K. Liu, T. Tan, W. Zhang, and X. Gao, “An acoustic model for microresonator in on-beam quartz-enhanced photoacoustic spectroscopy,” Appl. Phys. B 108(2), 361–367 (2012). [CrossRef]
5. H. Yi, K. Liu, S. Sun, W. Zhang, and X. Gao, “Theoretical analysis of off beam quartz-enhanced photoacoustic spectroscopy sensor,” Opt. Commun. 285(24), 5306–5312 (2012). [CrossRef]
6. K. Liu, H. Yi, A. A. Kosterev, W. Chen, L. Dong, L. Wang, T. Tan, W. Zhang, F. K. Tittel, and X. Gao, “Trace gas detection based on off-beam quartz enhanced photoacoustic spectroscopy: optimization and performance evaluation,” Rev. Sci. Instrum. 81(10), 103103 (2010). [CrossRef] [PubMed]
7. T. Rück, R. Bierl, and F. Matysik, “NO2 trace gas monitoring in air using off-beam quartz enhanced photoacoustic spectroscopy (QEPAS) and interference studies towards CO2, H2O and acoustic noise,” Sens. Actuators B Chem. 255(2), 2462–2471 (2018). [CrossRef]
8. J. C. Petersen, L. Lamard, Y. Feng, J. Focant, and M. Lassen, “Quartz-enhanced photoacoustic spectroscopy as a platform for non-invasive trace gas analyser targeting breath analysis,” arXiv 1704.07442 (2017).
9. H. Yi, W. Chen, S. Sun, K. Liu, T. Tan, and X. Gao, “T-shape microresonator-based high sensitivity quartz-enhanced photoacoustic spectroscopy sensor,” Opt. Express 20(8), 9187–9196 (2012). [CrossRef] [PubMed]
10. F. Vandenbussche, I. Vaseva, K. Vissenberg, and D. Van Der Straeten, “Ethylene in vegetative development: a tale with a riddle,” New Phytol. 194(4), 895–909 (2012). [CrossRef] [PubMed]
11. J. C. Pech, M. Bouzayen, and A. Latché, “Climacteric fruit ripening: Ethylene-dependent and independent regulation of ripening pathways in melon fruit,” Plant Sci. 175(1–2), 114–120 (2008). [CrossRef]
12. S. Schilt, A. A. Kosterev, and F. K. Tittel, “Performance evaluation of a near infrared QEPAS based ethylene sensor,” Appl. Phys. B Lasers Opt. 95(4), 813–824 (2009). [CrossRef]
13. Z. Wang, Z. Li, and W. Ren, “Quartz-enhanced photoacoustic detection of ethylene using a 10.5 μm quantum cascade laser,” Opt. Express 24(4), 4143–4154 (2016). [CrossRef] [PubMed]
14. Z. Wang, J. Geng, and W. Ren, “Quartz-Enhanced Photoacoustic Spectroscopy (QEPAS) detection of the v7 band of ethylene at low pressure with CO2 interference analysis,” Appl. Spectrosc. 71(8), 1834–1841 (2017).
15. Z. Loghmari, M. Bahriz, D. D. Thomas, A. Meguekam, H. N. Van, R. Teissier, and A. N. Baranov, “Room temperature continuous wave operation of InAs/AlSb-based quantum cascade laser at λ ∼11 µm,” Electron. Lett. 54(17), 1045–1047 (2018). [CrossRef]
16. R. Arndt, “Analytical line shapes for Lorentzian signals broadened by modulation,” J. Appl. Phys. 36(8), 2522–2524 (1965). [CrossRef]
17. I. E. Gordon, L. S. Rothman, C. Hill, R. V. Kochanov, Y. Tan, P. F. Bernath, M. Birk, V. Boudon, A. Campargue, K. V. Chance, B. J. Drouin, J.-M. Flaud, R. R. Gamache, J. T. Hodges, D. Jacquemart, V. I. Perevalov, A. Perrin, K. P. Shine, M.-A. H. Smith, J. Tennyson, G. C. Toon, H. Tran, V. G. Tyuterev, A. Barbe, A. G. Császár, V. M. Devi, T. Furtenbacher, J. J. Harrison, J.-M. Hartmann, A. Jolly, T. J. Johnson, T. Karman, I. Kleiner, A. A. Kyuberis, J. Loos, O. M. Lyulin, S. T. Massie, S. N. Mikhailenko, N. Moazzen-Ahmadi, H. S. P. Müller, O. V. Naumenko, A. V. Nikitin, O. L. Polyansky, M. Rey, M. Rotger, S. W. Sharpe, K. Sung, E. Starikova, S. A. Tashkun, J. V. Auwera, G. Wagner, J. Wilzewski, P. Wcisło, S. Yu, and E. J. Zak., “The HITRAN2016 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transf. 203, 3–69 (2017). [CrossRef]
18. Z. Li, C. Shi, and W. Ren, “Mid-infrared multimode fiber-coupled quantum cascade laser for off-beam quartz-enhanced photoacoustic detection,” Opt. Lett. 41(17), 4095–4098 (2016). [CrossRef] [PubMed]
19. A. Miklós, P. Hess, and Z. Bozóki, “Application of acoustic resonators in photoacoustic trace gas analysis and metrology,” Rev. Sci. Instrum. 72(4), 1937–1955 (2001). [CrossRef]
20. L. E. Kinsler, A. R. Frey, A. B. Coppens, and J. V. Sanders, Fundamentals of Acoustics, 4th Edition (Wiley-VCH, n.d.).
21. D. V. Serebryakov, I. V. Morozov, A. A. Kosterev, and V. S. Letokhov, “Laser microphotoacoustic sensor of ammonia traces in the atmosphere,” Quantum Electron. 40(2), 167–172 (2010). [CrossRef]
22. D. A. Bohn, “Environmental effects on the speed of sound,” in 83rd AES Convention (Audio Engineering Society, 1987), paper 2486.
23. N. Petra, J. Zweck, A. A. Kosterev, S. E. Minkoff, and D. Thomazy, “Theoretical analysis of a quartz-enhanced photoacoustic spectroscopy sensor,” Appl. Phys. B Lasers Opt. 94(4), 673–680 (2009). [CrossRef]
24. S. Schilt and L. Thévenaz, “Wavelength modulation photoacoustic spectroscopy: theoretical description and experimental results,” Infrared Phys. Technol. 48(2), 154–162 (2006). [CrossRef]
