Open Access
Add to BibSonomy
Abstract
A fast gas sensing scheme based on a non-dispersive frequency comb spectroscopy (ND-FCS) is proposed and experimentally demonstrated. Its capacity for multi-component gas measurement is experimentally investigated as well, by using the time-division-multiplexing (TDM) method to realize specific wavelength selection of the fiber laser optical frequency comb (OFC). A dual-channel optical fiber sensing scheme is established with a sensing path consisting of a multi-pass gas cell (MPGC), and a reference path with a calibrated signal to track the repetition frequency drift of the OFC for a real-time lock-in compensation and system stabilization. The long-term stability evaluation and the simultaneous dynamic monitoring are carried out, with the target gases of ammonia (NH3), carbon monoxide (CO) and carbon dioxide (CO2). The fast CO2 detection in human breath is also conducted. The experimental results show that at an integration time of 10 ms, the detection limits of the three species are evaluated to be 0.0048%, 0.1869% and 0.0467%, respectively. A low minimum detectable absorbance (MDA) down to 2.8 × 10−4 can be achieved and a dynamic response with millisecond time can be realized. Our proposed ND-FCS exhibits excellent gas sensing performance with merits of high sensitivity, fast response and long-term stability. It also shows great potential for multi-component gas monitoring in atmospheric monitoring applications.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Spectroscopic gas sensing technology plays a significant role in applications like chemical analysis [1–3], industrial monitoring [4,5] and medical diagnosis [6,7]. With the development of the laser source and the corresponding spectroscopic techniques, the gas detection system has been rapidly promoted towards high precision, multi-component sensitivity, fast response and systematical robustness. However, these features are hardly to be realized in one system. For instance, the most widely used tunable diode laser absorption spectroscopy (TDLAS) has merits of high sensitivity, high selectivity and fast response based on the narrowband coherent light source. With the restrictions on the wavelength scanning range, it needs multiple laser sources for multi-component gas detection. In contrast, gas sensing systems based on broadband incoherent light sources are suitable for multi-component gas measurements, while exhibits low sensitivity, poor stability and relatively low spectrum acquisition efficient.
Optical frequency comb (OFC), regarded as a revolutionary laser with both high coherence and wide spectral coverage [8–10], is particularly suitable for high-precision, broadband molecular spectroscopy. The ultra-fast pulse feature of the OFC guarantees the fast gas detection as well. Therefore, the frontier applications, for instance chemical reaction kinetics [11,12], transient combustion diagnosis [13,14] and multi-species breath analysis [15,16], have been extensively studied. In the past decade, the frequency comb spectroscopy (FCS) based on the spatially dispersive scheme [17–19], the Michelson-based Fourier transform [20,21] and the dual-comb configuration [22–24] have been proposed. The unprecedented performances of such FCS on precision, wide bandwidth and high time resolution for gas absorption spectroscopy have been experimentally demonstrated. These systems commonly rely on sophisticated optical dispersion or interferometric configurations for high-finesse gas absorption spectrum detection, and mostly utilize the mode-locked fiber laser as the OFC source to ensure the broadband spectral coverage. The long-term stability requires such lasers with actively mode-locking feedback and frequency stabilization system, which make the system complicated with high costs and expenses.
Here, we propose a non-dispersive [25,26] frequency comb spectroscopy (ND-FCS) with fast sensing speed based on a free-running laser frequency comb and a tunable optical filter (TOF). The multi-component gas measurement is realized and experimentally demonstrated. The novel aspects of the ND-FCS are: (1) The intensity demodulation method by adopting a radio frequency beating and a lock-in amplification is utilized based on the high-speed pulse feature of OFC, guaranteeing the high sensitivity (via 1/f noise suppression), high measuring speed (fast response) and simple structure. (2) The reference channel with the radio frequency beat note feedbacks the first harmonic (1f) lock-in reference signal, providing the background intensity noise and the repetition frequency drift of the laser pulse. The long-term stability of the system can be greatly improved by real-time intensity noise suppression and frequency drift compensation even by using a low-cost free running laser frequency comb with simple configuration. (3) Multi-component gas measurement is experimentally demonstrated, by programmatic tuning of the TOF to fast select specific narrow wavebands, via the time-division-multiplexing (TDM) operation. Benefiting from the high coherent light signals, the high sensitivity and fast response for each gas species can be remained.
The following of the paper consists of 3 parts. Firstly, the principles and system setup will be elaborated. Then the experiments and results including the system calibration and sensitivity test, the stability and detection limit evaluation, the multi-gas dynamic measurement and the human breath detection are presented in details. Finally, the conclusion of the work is given and the potential applications will be discussed.
2. Principles and system setup
2.1 ND-FCS gas sensing theory
The broadband laser pulse from the OFC is selected by a narrow bandpass optical filter to generate a narrowband sensing laser pulse covering a specific gas absorption line. The sensing pulse passes through a multi-pass gas cell (MPGC) and its intensity is detected for determination of gas concentration according to the Beer-Lambert law. The intensity of the transmitted sensing laser pulse $I({{\lambda_1}} )$ can be written as
where ${I_0}({{\lambda_1}} )$ is the original intensity of the sensing laser pulse and $\alpha ({{\lambda_1}} )$ is the gas absorption coefficient at the peak wavelength ${\lambda _1}$. $\gamma $ is the overlap coefficient between the narrow bandwidth of the sensing laser and the gas absorption line. Here the product of $\alpha ({{\lambda_1}} )$ and $\gamma $ represents the integration of the absorption coefficient over the transmission band of the TOF, C and L represent the gas concentration and the light-gas interaction length, respectively.To suppress the intensity noise, a dual-channel optical path is established with two photodetectors (PDs) converting the laser pulses to electrical signals. Their intensities are ${U_\textrm{d}}({{\lambda_1}} )$ and ${U_\textrm{r}}({{\lambda_1}} )$ for the detection channel and the reference channel, respectively. Taking the propagation loss $\varepsilon $ and the photodetection coefficient $\mu $ into consideration with the subscripts $\textrm{d}$ and $\textrm{r}$ identifying the channels, ${U_\textrm{d}}({{\lambda_1}} )$ and ${U_\textrm{r}}({{\lambda_1}} )$ can be expressed as
To measure the intensity of a pulsed signal with radio repetition frequency ${f_\textrm{r}}$, electronical frequency mixing is firstly applied for frequency down-conversion. The generated beat note ${f_\textrm{i}}$ is then lock-in amplified to extract a DC signal (in V) with gain factors of ${K_\textrm{d}}$ and ${K_\textrm{r}}$ respectively. The finally acquired detection signal ${S_\textrm{d}}({{\lambda_1}} )$ and the reference signal ${S_\textrm{r}}({{\lambda_1}} )$ can be expressed as
Normalization of the detection signal is conducted as ${S_\textrm{d}}({{\lambda_1}} )/{S_\textrm{r}}({{\lambda_1}} )$, which corresponds to the gas concentration C according to the relation
By calibration using various samples of the target gas, the constant terms $\textrm{ln}({{K_\textrm{d}}{\varepsilon_\textrm{d}}{\mu_\textrm{d}}/{K_\textrm{r}}{\varepsilon_\textrm{r}}{\mu_\textrm{r}}} )$ and $1/\alpha ({{\lambda_1}} )\gamma L$ can be determined as ${x_1}$ and ${y_1}$, respectively. The equation is simplified as
2.2 System setup
A schematic diagram of the ND-FCS system is shown in Fig. 1. The OFC source is homemade based on a fiber laser frequency comb without any active stabilizing configuration. Emitting at center wavelength of 1560 nm with a bandwidth of ∼ 30 nm, it has an output power of ∼ 10 mW and a pulse repetition rate of ∼ 41.708 MHz. A TOF (WL Photonics, Model MPTF-NE-S-1550-110/0.1-SM-0.9/1.0-FC/APC-USB) with a bandwidth of 0.1 nm is used to select specific modes from the OFC source. The obtained narrow laser pulse is divided equally into a reference signal and a sensing probe, while a tiny fraction is extracted for wavelength monitoring via an optical spectrum analyzer (OSA, YOKOGAWA, Model AQ6370D). A MPGC with a reflective length of ∼ 4 m serves as the light-gas interaction path, which results from 38 reflections between two 1-inch mirrors in a cell body sized 126 × 34 × 34 mm3. The reference channel has a tunable optical attenuator (TOA, Micro photons, Model VOA-W1550-1-5-S15A) to adjust the intensity of the background signal. Two PDs (Conquer Optics, Model KG-APR-1G-A-FS-AC) of the same model receive the dual-channel laser pulses and output the corresponding electrical pulses. The optical path is all fiber interconnected, ensuring the systematical conciseness and stability.
Fig. 1. Schematic diagram of the ND-FCS gas sensing system consisting of an optical part and an electrical part with dual-channel configuration. OSA: optical spectrum analyzer; MPGC: multi-pass gas cell; TOA: tunable optical attenuator; DAQ: data acquisition; FFT: fast Fourier transform; LIA: lock-in amplifier.
A dual-channel signal generator (Siglent, Model SDG6052X-E) with a bandwidth of 500 MHz generates two radio frequencies of ∼ 41.758 MHz as local oscillators. Two electronical frequency mixers (Pasternack, Model PE86X1018) are employed to beat the outputs of the PDs with the radio frequencies. As a result, the original repetition frequency of the laser pulse is down-converted to the intermediate frequency of ∼ 50 kHz. A data acquisition (DAQ) card (National Instruments, Model USB-6361) with a sampling rate of 1 MHz performs analog-to-digital conversion of the beating signals. Based on a self-developed LabVIEW platform, a digital dual-channel lock-in amplifier (LIA1 and LIA2) measures the amplitudes of the two beat notes. The fast Fourier transform (FFT) is conducted to calculate the actual frequency of the beat note from the reference channel. It then feedbacks to the LIAs as the 1f reference signal to compensate the repetition frequency drift of the OFC in real time. The time constant of the lock-in amplification is set as 1 ms for gas concentration calculation. Considering higher signal-to-noise ratio (SNR), data averaging window of 10 ms is configured. In other words, the time resolution for gas detection is consequently 10 ms. The software platform reads/writes status/controlling commands through serial communication, to tune the working wavelengths of the TOF and change the working wavelength of the sensing probe.
2.3 Lock-in frequency drift compensation scheme
The real-time beat note monitoring and lock-in reference compensation process is tested by 30-min continuous measurements of the OFC emission intensity transmitted in the sensing channel without the MPGC. As shown in Fig. 2, both two sets of the monitored beat frequency reveal the drifts of the OFC repetition frequency up to ∼ 30 Hz in free-running operation. Accordingly, the measured pulse intensity plotted by the blue curve in Fig. 2(a) undergoes a severe variation up to 0.104 V, when the 1f lock-in reference is fixed at 50.246 kHz. In comparison, a real-time frequency drift compensation is realized by feeding the monitored beat frequency to the lock-in amplification as the 1f reference signal. Figure 2(b) shows great improvement of the stability with the measured intensity variance in a tiny range of 0.008 V. Consequently, the stability of the ND-FCS is guaranteed by using the lock-in compensating method.
Fig. 2. Measured laser pulse intensity and the monitored beat frequency of the OFC in 30-min experiment without and with the proposed lock-in compensation scheme are plotted in (a) and (b), respectively.
2.4 TOF configuration
The spectrum of the employed laser frequency comb is shown in Fig. 3(a). In the waveband, gases including ammonia (NH3), carbon monoxide (CO) and carbon dioxide (CO2) exhibit strong absorption features as shown in Fig. 3(b). The absorption coefficient profiles are obtained by numerical simulation based on the HITRAN database [27] at 1 atm, 296 K and 100% concentration. Considering the higher power of the sensing pulse, the less mutual interference between different gases and the higher absorption coefficient of the gas molecule, the sensing pulse is selected for each gas to cover the specific absorption lines. The center wavelengths of the TOF are accordingly set at 1548.9 nm, 1567.3 nm and 1572.3 nm.
Fig. 3. (a) Output spectrum of the OFC measured using OSA. (b) Absorption spectra of NH3, CO and CO2 with pure density in the bandwidth of the OFC. (c), (d) and (e) The overlap condition between the gas absorption lines and the bandwidth of the tunable filter, at wavelengths of 1548.9 nm, 1567.3 nm and 1572.3 nm, respectively.
Figure 3(c), (d) and (e) present the overlap condition between the gas absorption lines and the bandwidth of the TOF, at wavelengths of 1548.9 nm, 1567.3 nm and 1572.3 nm, respectively. It can be clearly seen in Fig. 3(b) that the NH3 absorption suffers negligible interference from CO and CO2, thus potential cross-interference is depicted only in Fig. 3(d) and (e). According to the transition overlapping condition, NH3 absorption does contribute to the lines of CO and CO2 when the concentrations of them are on the same order of magnitude. But this cross-interference can be effectively suppressed by setting the NH3 dynamic range an order of magnitude lower than that of the other two species. The measured bandwidth of the TOF is nearly 0.085 nm, which is much narrower than the complex combination line of NH3, yet comparable to the lines of CO and CO2. As a result, higher sensitivity on absorbance detection for NH3 is believed to be achieved due to the higher overlap coefficient $\mathrm{\gamma }$ as described by Eq. (1). Moreover, the power of sensing probes for NH3, CO and CO2 are measured to be 9 µW, 5 µW and 11 µW, respectively. Thus, the sensitivity for CO absorbance detection may be lower than that of the other two gases due to the poorer SNR. The center wavelength of the TOF can be continuously switched with an interval of 1 s to cover different gas absorption lines, which makes it possible for quasi synchronous multi-component gas analysis.
3. Results and discussions
3.1 System calibration and sensitivity estimation
Based on the selected absorption lines for the three gas samples, calibration of the ND-FCS system is performed within the NH3 concentration range of 0% - 1%, the CO concentration range of 0% - 20% and the CO2 concentration range of 0% - 10%. For each gas, seven standard samples of different concentrations are prepared via quantitative dilution with nitrogen (N2) of 5N purity (99.999% in volume fraction). The generated samples are successively introduced into the MPGC. Each concentration measurement lasts for 3 min after the stabilization of the output signal. Figure 4(a), (b) and (c) plot the actual sample concentration versus the averaged signal amplitude over each 3-min measurement for NH3, CO and CO2, respectively. The original recorded data with a sampling time interval of 10 ms is also given in the insets. The standard deviation of the data set for each concentration is adopted to provide error bar, so that the measuring uncertainty can be characterized. The set concentrations, averaged signal amplitudes and standard deviations of the measured data sets are listed in Table 1 for reference. By exponentially fitting on the plotted data points, calibration formulas for the three gases are obtained and written as
Fig. 4. Different gas concentrations versus the measured normalized signal amplitudes and their logarithmic fitting curve for (a) NH3, (b) CO and (c) CO2, respectively. The insets depict the original sampling results of the three gases with different concentrations at integration time of 10 ms.
Table 1. Set concentrations, averaged normalized signal amplitudes and standard deviations in the calibration.
It should be mentioned that the constant value -0,01119, -0.39579 and -0.18768 are the detection uncertainties induced from the experimental noise.
In the fitting process, R2 (covariance of deviation) values of 0.99989, 0.99904 and 0.99898 are achieved for NH3, CO and CO2, respectively. The apparently nonlinear behavior of the CO curve is caused by the wider concentration range with the corresponding absorbance range of 0 - 0.2. While, for NH3 and CO2, the involved absorbance range are both about 0 - 0.06, leading to nearly linear calibration relation. Even though the CO absorption coefficient is about 2 times of CO2 according to Fig. 3(b), the much lower sensitivity for CO is clearly observed by comparison of Fig. 4(b) and (c). It is deduced to the much lower power of the sensing probe for CO than that of the other two components, which leads to the lower SNR in signal detection. The error bars statistically indicate the signal fluctuations and the detection uncertainties for NH3, CO and CO2 are estimated to be 0.0062%, 0.279% and 0.077%, respectively, by calculating the averaged standard deviation of the seven data sets for each species.
3.2 Stability and detection limit evaluation
To further evaluate the stability of the developed system and quantitatively characterize the detection limit performance, the long-term continuous measurement of a single sample is carried out. Considering that the dynamic gas diluting process may introduce concentration fluctuation, N2 gas samples is under test, named zero-gas experiments [28] and for each gas species the TOF is tuned to realize the corresponding wavelength selection. The measured concentration values during an operation of 30 min for NH3, CO and CO2 are presented in Fig. 5(a), (b) and (c), respectively. Good stability is observed in all the three sets of results, among which the data for NH3 exhibits the lowest and steadiest noise distribution.
Fig. 5. Measured concentrations of (a) NH3, (b) CO and (c) CO2 in stability assessing experiments in the duration of 30 min. Accordingly, Allan deviation curves are further calculated and plotted in (d), (e) and (f) to analyze the detection limits.
At the integration time of 10 ms, a segment of 10 min from the original data set is extracted to calculate Allan deviation distribution for each gas under test. The detection limit and noise feature of the system are analyzed. As shown in Fig. 5(d), the deviation curve of NH3 fits well with the line proportional to $1/\textrm{sqrt}\left( \mathrm{\tau } \right)$ until 20 s, which indicates that the sensor affected dominantly by the White-Gaussian noise. As for CO and CO2, the evaluated detection limit shown in Fig. 5(e) and (f) both reaches the minimum at ∼ 1.2 s. It can also be verified by the time domain fluctuations in Fig. 5(b) and (c), which obviously include not just high-frequency random noise. Based on the 1σ deviation value at 10 ms, the detection limits for NH3, CO and CO2 are obtained as 0.0048%, 0.1869% and 0.0467%, respectively. The corresponding absorbance detection limits are 2.88 × 10−4, 2.02 × 10−3 and 2.8 × 10−4, respectively. The one order of magnitude larger value for CO here is due to the poor SNR of the sensing laser probe as well. Moreover, the optimum detection limits for the three components can be estimated to be 1 part per million (ppm), 190 ppm and 87 ppm respectively at 20 s, 1.2 s and 1.2 s.
The minimum detectable absorbance (MDA) of the proposed ND-FCS is comparable to that of the wavelength modulation spectroscopy (WMS) [28–30], which is a well-developed gas sensing technique using tunable narrow-linewidth CW lasers. Additionally, our system shows much higher detection speed and efficiency compared with these systems [28–30], benefiting from the fast pulses of OFC and the non-dispersive scheme without the wavelength scanning operation.
3.3 Multi-gas dynamic measurement
The multi-component gas sensing capacity of the developed ND-FCS system is validated by dynamic measurement of a three-gas mixture. A four-channel gas mixing system is used for the mixture sample preparation, with an output flow rate of 100 standard cubic centimeters per minute (sccm). The prepared sample is continuously introduced into the MPGC, and the constituent concentrations are changed every 5 min. Figure 6 shows the measured concentrations of the three components, by using the TDM scheme with a TOF switching interval of 1 s. Delaying 0.5s after tuning the TOF for stabilization, gas concentration is then recorded with an integration time of 10 ms and lasts for 0.5s for each gas. Consequently, 50 detected values for each gas species are output at a circulating interval of 3 s. The actual species concentrations, i.e. the setpoints, are marked above the measured data points. From the response curves, one can see that each gas of the mixture experiences five different concentration evolutions, including a rapid change period and a stable sensing duration. The response time of the system for rising or falling processes is observed to be ∼ 120 s, which is relatively long due to the actual low diffusion rate of the gas to filling the MPGC with volume of ∼ 50 mL.
Fig. 6. Measured concentrations of NH3, CO and CO2 in gas mixture with a TDM switching interval of 1 s and an integration time of 10 ms.
For each gas sample, the average values of the determined concentrations with error bars are plotted versus the setpoints in Fig. 7. The data is also listed in Table 2, with the measurement uncertainties provided as standard deviations (σ). By the linear fitting, the detection linearities up to 0.99996, 0.99894 and 0.99967 are obtained for NH3, CO and CO2, respectively, indicating that the measured dynamic results coincide well with the set values. Averaging the five standard deviations for each species, the cross-sensitivities for NH3, CO and CO2 are calculated to be 0.003%, 0.335% and 0.085%, respectively. These values are on the same orders of magnitude with that obtained in Section 3.1, indicating that the sensitivity performance suffers no deterioration in real-time multi-component gas detection. It is benefit from the negligible interference between the absorption lines of the three species, and the TDM scheme resulting no cross-disturbance between the three waveband channels. It should be mentioned that the measured concentration drift occurs in several sensing durations, which can be clearly seen in Fig. 6, for example for carbon dioxide from 5 min on and for carbon monoxide from 10 min on. From Table 2, one can also find that the uncertainties of the corresponding measured concentrations are relatively larger. This phenomenon perhaps results from the differential-mode gain variation of the two PDs and the temperature-induced deformation of the light path in the MPGC, which both lead to accidental and slow change of the detection signal intensity. It is believed that utilizing balanced detectors and introducing temperature controllers for the MPGC could improve the stability of the system.
Fig. 7. Concentration values obtained in the multi-gas dynamic measurement versus the actual concentrations and their linear fitting curve for (a) NH3, (b) CO and (c) CO2, respectively.
Table 2. Setpoints and measured concentrations in the multi-gas measurement
3.4 Human breath detection
Human breath detection is carried out to test the sensor performance on the repeatability and fast response. To introduce the exhaled gas into the MPGC, a tee joint equipped with a check valve is built as the human breath sampling port and a N2 supply with high flow rate of 1000 sccm acts as a boosting channel to improve the filling time of the gas (see the inset of Fig. 8). The breath is preprocessed by using magnesium sulfate-based desiccant to remove the water content. Ten breath tests are conducted with random exhalation strength. Figure 8 presents the monitoring results of the CO2 concentration, during the 85-second experimental process, in which 10 different breath processes are clearly recorded with distinct intensities and time durations. The maximum CO2 concentration is observed to be ∼ 5%, while the minimum value is ∼1%. The overall stable baseline reveals good repeatability of the system. The baseline fluctuations and the negative concentration impact may be due to the fast pressure variations or slight optical path deformations in the MPGC. Among the captured breath responses, the slightest one at the time of ∼ 48 s is enlarged and the detailed process is depicted in the inset of Fig. 8. With the time resolution of 10 ms, the concentration variances are clearly displayed. The rising time and falling time are both observed to be ∼ 200 ms, which indicates the possibility and potential for fast gas monitoring applications.
Fig. 8. Dynamic measurement of the CO2 concentration during the human breath test with an integration time of 10 ms. The insets are the photograph of the breath sampling port and the detailed response curve to a slight breath.
4. Conclusion
The ND-FCS for fast gas sensing and multi-component gas measurement based on a free-running fiber laser frequency comb and a tunable narrow-linewidth optical filter is developed and experimentally demonstrated. A lock-in configuration with real-time frequency feedback is established to compensate the repetition frequency drifts of the OFC. System calibration within a NH3 concentration range of 0% - 1%, a CO concentration range of 0% - 20% and a CO2 concentration range of 0% - 10% is carried out. Through long-term measurements and Allan deviation analyses, the detection limits on NH3, CO and CO2 at integration time of 10 ms are obtained to be 0.0048%, 0.1869% and 0.0467%, respectively. An absorbance detection sensitivity up to 2.8 × 10−4 is consequently achieved. Using TDM, the quasi synchronous dynamic measurement of the three gases is realized. Human breath test validates the high performance of the system on fast CO2 monitoring, according to the clear response to the concentration variation at 200-ms time scale.
The proposed ND-FCS exhibits excellent sensing ability, without the needs of low-efficiency spectrum scanning and complex laser mode locking configuration, providing a promising candidate for long-term, high sensitivity, fast response gas sensing applications and multi-component measurement requirements.
Funding
Center-initiated Research Project of Zhejiang Lab (2022ME0AL03); National Natural Science Foundation of China (62205301).
Acknowledgements
The authors wish to express their gratitude to the Center-initiated Research Project of Zhejiang Lab (No. 2022ME0AL03) and the National Natural Science Foundation of China (No. 62205301), for their supports to this work. Also, thanks to Prof. Yunjiang Rao, Prof. Baicheng Yao and Dr. Teng Tan, for their assistance with the design and fabrication of the fiber laser frequency comb.
Disclosures
The authors declare no conflicts of interest.
Data availability
The data that support the results of this work are available from the authors upon reasonable request
References
1. Y. B. Ji, K. Duan, Z. M. Lu, and W. Ren, “Mid-infrared absorption spectroscopic sensor for simultaneous and in-situ measurements of ammonia, water and temperature,” Sens. Actuators, B 371, 132574 (2022). [CrossRef]
2. Z. W. Liu, C. T. Zheng, T. Y. Zhang, Y. F. Li, Q. Ren, C. Chen, W. L. Ye, Y. Zhang, Y. D. Wang, and F. K. Tittel, “Midinfrared sensor system based on tunable laser absorption spectroscopy for dissolved carbon dioxide analysis in the South China Sea: System-level integration and deployment,” Anal. Chem. 92(12), 8178–8185 (2020). [CrossRef]
3. Z. W. Liu, C. T. Zheng, T. Y. Zhang, Y. Zhang, Y. D. Wang, and F. K. Tittel, “High-precision methane isotopic abundance analysis using near-infrared absorption spectroscopy at 100 Torr,” Analyst 146(2), 698–705 (2021). [CrossRef]
4. X. T. Lou, Y. B. Feng, C. Chen, and Y. K. Dong, “Multi-point spectroscopic gas sensing based on coherent FMCW interferometry,” Opt. Express 28(6), 9014–9026 (2020). [CrossRef]
5. T. Zhang, Y. X. Xing, G. X. Wang, and S. L. He, “High sensitivity continuous monitoring of chloroform gas by using wavelength modulation photoacoustic spectroscopy in the near-infrared range,” Appl. Sci. 11(15), 6992 (2021). [CrossRef]
6. M. Azhar, J. Mandon, A. H. Neerincx, Z. Liu, J. Mink, P. J. F. M. Merkus, S. M. Cristescu, and F. J. M. Harren, “A widely tunable, near-infrared laser-based trace gas sensor for hydrogen cyanide (HCN) detection in exhaled breath,” Appl. Phys. B 123(11), 268 (2017). [CrossRef]
7. R. Q. Yu, H. Xia, T. Pang, B. Wu, Z. Li, P. S. Sun, Q. Guo, Q. Y. Li, and Z. R. Zhang, “The measurement of delta C-13 in respiratory gases based on incoherent broadband cavity enhanced absorption spectroscopy,” Infrared Phys. Technol. 117, 103812 (2021). [CrossRef]
8. T. Udem, R. Holzwarth, and T. W. Hansch, “Optical frequency metrology,” Nature 416(6877), 233–237 (2002). [CrossRef]
9. 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]
10. S. T. Cundiff and J. Ye, “Colloquium: femtosecond optical frequency combs,” Rev. Mod. Phys. 75(1), 325–342 (2003). [CrossRef]
11. F. C. Roberts, H. J. Lewandowski, B. F. Hobson, and J. H. Lehman, “A rapid, spatially dispersive frequency comb spectrograph aimed at gas phase chemical reaction kinetics,” Mol. Phys. 118(16), 1733116 (2020). [CrossRef]
12. M. A. Abbas, L. Van Dijk, K. E. Jahromi, M. Nematollahi, F. J. M. Harren, and A. Khodabakhsh, “Broadband time-resolved absorption and dispersion spectroscopy of methane and ethane in a plasma using a mid-infrared dual-comb spectrometer,” Sensors 20(23), 6831 (2020). [CrossRef]
13. N. H. Pinkowski, Y. M. Ding, C. L. Strand, R. K. Hanson, R. Horvath, and M. Geiser, “Dual-comb spectroscopy for high-temperature reaction kinetics,” Meas. Sci. Technol. 31(5), 055501 (2020). [CrossRef]
14. P. J. Schroeder, A. S. Makowiecki, M. A. Kelley, R. K. Cole, N. A. Malarich, R. J. Wright, J. M. Porter, and G. B. Rieker, “Temperature and concentration measurements in a high-pressure gasifier enabled by cepstral analysis of dual frequency comb spectroscopy,” Proc. Combust. Inst. 38(1), 1561–1569 (2021). [CrossRef]
15. F. Keilmann and S. Amarie, “Mid-infrared frequency comb spanning an octave based on an Er fiber laser and difference-frequency generation,” J. Infrared, Millimeter, Terahertz Waves 33(5), 479–484 (2012). [CrossRef]
16. Q. Z. Liang, T. C. Chan, P. B. Changala, D. J. Nesbitt, J. Ye, and J. Toscano, “Ultrasensitive multispecies spectroscopic breath analysis for real-time health monitoring and diagnostics,” Proc. Natl. Acad. Sci. USA118(40), e2105063118 (2021). [CrossRef]
17. N. Picque and T. W. Hansch, “Frequency comb spectroscopy,” Nat. Photonics 13(3), 146–157 (2019). [CrossRef]
18. G. Kowzan, D. Charczun, A. Cygan, R. S. Trawinski, D. Lisak, and P. Maslowski, “Broadband optical cavity mode measurements at Hz-level precision with a comb-based VIPA spectrometer,” Sci. Rep. 9(1), 8206 (2019). [CrossRef]
19. A. Klose, G. Ycas, F. C. Cruz, D. L. Maser, and S. A. Diddams, “Rapid, broadband spectroscopic temperature measurement of CO2 using VIPA spectroscopy,” Appl. Phys. B 122(4), 78 (2016). [CrossRef]
20. R. Dubroeucq and L. Rutkowski, “Optical frequency comb Fourier transform cavity ring-down spectroscopy,” Opt. Express 30(8), 13594–13602 (2022). [CrossRef]
21. I. Sadiek, T. Mikkonen, M. Vainio, J. Toivonen, and A. Foltynowicz, “Optical frequency comb photoacoustic spectroscopy,” Phys. Chem. Chem. Phys. 20(44), 27849–27855 (2018). [CrossRef]
22. M. Soriano-Amat, M. A. Soto, V. Duran, H. F. Martins, S. Martin-Lopez, M. Gonzalez-Herraez, and M. R. Fernandez-Ruiz, “Common-path dual-comb spectroscopy using a single electro-optic modulator,” J. Lightwave Technol. 38(18), 5107–5115 (2020). [CrossRef]
23. J. Chen, K. Nitta, X. Zhao, T. Mizuno, T. Minamikawa, F. Hindle, Z. Zheng, and T. Yasui, “Adaptive-sampling near-Doppler-limited terahertz dual-comb spectroscopy with a free-running single-cavity fiber laser,” Adv. Photonics 2(03), 036004 (2020). [CrossRef]
24. G. Ycas, F. R. Giorgetta, E. Baumann, I. Coddington, D. Herman, S. A. Diddams, and N. R. Newbury, “High-coherence mid-infrared dual-comb spectroscopy spanning 2.6 to 5.2 µm,” Nat. Photonics 12(4), 202–208 (2018). [CrossRef]
25. L. B. Zhou, Y. Y. He, Q. Zhang, and L. Zhang, “Carbon dioxide sensor module based on NDIR technology,” Micromachines 12(7), 845 (2021). [CrossRef]
26. M. Vafaei and A. Amini, “Chamberless NDIR CO2 sensor robust against environmental fluctuations,” ACS Sens. 6(4), 1536–1542 (2021). [CrossRef]
27. C. Hill, I. E. Gordon, R. V. Kochanov, L. Barrett, J. S. Wilzewski, and L. S. Rothman, “HITRANonline: An online interface and the flexible representation of spectroscopic data in the HITRAN database,” J. Quant. Spectrosc. Radiat. Transfer 177, 4–14 (2016). [CrossRef]
28. Z. W. Liu, C. T. Zheng, C. Chen, Y. F. Li, H. T. Xie, Q. Ren, Y. D. Wang, and F. K. Tittel, “ICL-based mid-infrared carbon dioxide sensor system for deep-sea natural gas hydrate exploration,” Opt. Express 27(4), 5598–5609 (2019). [CrossRef]
29. C. G. Li, L. Dong, C. T. Zheng, and F. K. Tittel, “Compact TDLAS based optical sensor for ppb-level ethane detection by use of a 3.34 mu m room-temperature CW interband cascade laser,” Sens. Actuators, B 232, 188–194 (2016). [CrossRef]
30. Y. J. Du, Z. M. Peng, and Y. J. Ding, “Wavelength modulation spectroscopy for recovering absolute absorbance,” Opt. Express 26(7), 9263–9272 (2018). [CrossRef]
