Open Access
Add to BibSonomy
Abstract
In this work, I demonstrate a novel configuration of a photothermal gas sensor. Detection of nitric oxide at a wavelength of 5.26 µm was possible by constructing an absorption cell based on a self-fabricated antiresonant hollow core fiber characterized by low losses at both the pump and probe wavelengths. Proper design of the sensor allowed using the heterodyne interferometry-based signal readout of the refractive index modulation, which yielded a record noise equivalent absorption of 2.81×10−8 cm-1 for 100 s integration time for mid-infrared fiber-based gas sensors. The obtained results clearly demonstrate the full potential of using properly designed antiresonant hollow core fibers in combination with sensitive gas detection methods.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The incorporation of hollow-core fibers (HCF) in gas sensors is currently being intensively investigated by the laser spectroscopy community. Historically, such configurations were limited to the molecules having absorption spectra in the sub-2 µm wavelength region, mainly due to high silica absorption at longer wavelengths. This drawback was mitigated with the discovery and further development of fibers with limited material losses, namely, a type of HCFs – the antiresonant hollow-core fibers (ARHCF). In ARHCFs, the unique guiding mechanism allows achieving a minimum overlap between the glass capillaries forming the fiber cladding and the transmitted light (∼10−5). Properly designed ARHCFs can exhibit losses in the range of ∼4 dB/m at wavelengths up to 5.5 µm [1]. The laser spectroscopy community has quickly reacted to this breakthrough and proposed several unique sensor configurations, which took advantage of the new types of HCFs. Researchers designed and tested several dissimilar configurations of the sensors in which the HCFs were used as gas absorption cells (GAC), this includes traditional detection methods like tunable laser absorption spectroscopy (TDLAS) [2], wavelength modulation spectroscopy (WMS) [3–5], or techniques where the spectroscopy signal is encoded in an induced refractive index (RI) modulation – photothermal spectroscopy (PTS) [6–9]. This scheme of gas sensing can yield exceptionally good sensitivity with proper selection of the HCF and signal retrieval. In PTS, the gas molecules are excited (heated) by a pump laser matching the transitions of the targeted analyte, which results in localized variations of the gas refractive index (RI). The resulting RI modulation is miniscule, therefore several methods of readout were presented, mostly relying on Fabry-Perot or Mach-Zehnder interferometers [6–9]. In PTS, the magnitude of the RI modulation depends on a number of factors: intensity of the pump, gas concentration, absorption distance, overlap of pump and probe beams, pump modulation frequency, gas temperature, pressure and humidity, assuming other parameters are kept constant [10,11]. One of the most significant advantages of the PTS approach to gas detection lies in the fact that the pump and probe laser sources can be completely independent. This allows separating the two functions in the sensor and designing its layout to obtain a relatively noncomplex, yet selective and sensitive configuration. Currently, researchers focus on using PTS sensors employing quantum cascade lasers (QCLs) or interband cascade lasers (ICLs) as the pump, while keeping the probe part of the sensor based on inexpensive, reliable and compact fiber-based 1.55 µm technology. Up to the development of novel types of ARHCFs, the mid-IR PTS sensors were limited to severely more complex bulk optics-based configurations [12,13], or to HCFs with transmission characteristics limiting the sensor to sub-2.4 µm wavelength pump operation [14,15]. The especially interesting result was obtained for a PTS sensor working in the near-infrared (near-IR) wavelength region [16]. Authors were able to couple the light into the HCF and achieve LP01 and LP11 mode transmission for the pump and probe beams. This allowed them to use the mode-phase difference detection technique, which due to the common noise cancelation can reach exceptionally low noise equivalent absorption (NEA) – 1.6 × 10−11 cm-1. However, the sensor was limited to the sub-2 µm wavelength region, thus rendering it unsuitable for applications requiring mid-IR pump lasers for detection.
The published experimental results prove, that the advantages of using ARHCFs in gas sensors come from the tight confinement of the pump and the probe beams inside a hollow core, which can be subsequently conveniently filled with a gas sample under test, forming a robust and low-volume GAC. With proper sensor design, such detectors can achieve noise equivalent absorption limits outperforming traditional configurations, e.g. based on multipass cells, at the same time providing extremely low sample volumes (in the range of µl), and requiring noncomplex bulk-optic beam coupling [17].
In this work, I present a gas sensor employing a self-fabricated ARHCF-based GAC. Appropriate design and fabrication of the fiber resulted in its low losses in two crucial transmission windows – 1.55 µm (probe) and 5.26 µm (pump). This unique feature allowed me to employ the PTS detection technique and conveniently separate the pump laser and the probe laser in the experiment and target strong absorption lines of nitric oxide (NO) at a wavelength of 5.26 µm. To simplify the sensor configuration and omit the active interferometer stabilization requirements, a heterodyne PTS signal readout was employed, reaching a NEA equal to 2.86×10−7cm-1 and 2.81×10−8 cm-1, for 1 s and 100 s integration time, respectively. This is equal to a detection limit of 1.78 parts-per-billion-by-volume (ppbv) for 100 s integration time, utilizing only a 25 cm-long GAC.
Up to my best knowledge, the presented results set a record in two categories. Firstly, this is the first demonstration of combining a heterodyne PTS gas detection technique with a fiber-based GAC capable of detecting molecules beyond the 4.7 µm wavelength region. Secondly, the obtained NEA limit is the lowest reported for any fiber-based sensor capable of targeting gas molecules in the mid-IR. The presented results clearly show the perspective in the development of ARHCF-aided noncomplex and miniaturized gas detection platforms operating in the mid-IR wavelength region.
2. Sensor setup
The schematic of the sensor setup is depicted in Fig. 1. The main part of the sensor is the custom-built GAC. The GAC is constructed based on a 25 cm-long self-fabricated ARHCF with ∼4 dB/m and 6 dB/m attenuation at the pump and probe wavelength, respectively. More details about the fiber can be found in my previous work [3]. Femtosecond laser micromachining was used to side-drill holes in the middle part of the fiber, which enabled directly accessing the air core region. The total cross section of the holes is ∼ 0.0675 mm2, which is more than twice the cross section of the facet of the ARHCF. The processed fiber was inserted into a 1/16” T-joint from which two 8 cm long 1/16” stainless steel tubes protruded. A gas-tight connection was realized by UV-gluing the fiber to the ends of the steel tubes, as depicted in the schematic. This unique configuration provides a noncomplex gas sample delivery, fast purging of the gas cell, and does not require building air-tight gas cells on both ends of the ARHFC. This is especially important if a limited number of bulk optic components is the optimization goal for the sensor, or the signal retrieval method requires placing the beam coupling optics close to the fiber facet, as in this particular experimental setup.
Fig. 1. Schematic of the sensor setup. RF – RF spectrum analyzer, A- RF amplifier, DET – InGaAs detector, CPL – fiber coupler, AOM – acousto-optical modulator, QCL – quantum cascade laser, FL – focusing lens, G – germanium wedge, M – silver mirror, CL – fiber collimator, 1550–1550 nm laser, LDTC – laser driver with temperature controller, FG – function generator, GM – gas mixer. The inset shows a SEM image of the ARHCF with the fabricated microchannels.
The gas was delivered to the GAC via 1/4“ Teflon tubes directly from a commercial gas dilution system (Environics 4000), which allowed setting the desired gas concentration and flow through the sensor. A mid-IR QCL laser (Thorlabs, QD5500CM1) was used as the pump source. Its emission was temperature tuned to 10.38°C to reach the R6.5 transition of NO located at 1900.09 cm-1. The function generator was connected to the analog modulation input of the laser driver (Thorlabs, ITC4002QCL), which was used to modulate the wavelength of the laser with a sinusoidal waveform at a frequency f0 and a slow sawtooth ramp. The emission of the QCL was collimated, focused (calcium-fluoride lens, f=75 mm), and coupled into the ARHCF-based GAC with a transmission efficiency of 80% (82 mW of optical power was registered at the output of the GAC when filled with N2). An antireflection coated germanium wedge (G) was inserted in the beam path of the mid-IR beam to allow subsequently separating the near-IR beam exiting the GAC (see Fig. 1). Whereas the probe part of the sensor is constructed in a heterodyne detection configuration based on a near-infrared (near-IR) 1.55 µm telecom fiber technology, similar to my previous bulk optics-based experiments [12,18]. A distributed feedback (DFB) laser diode emitting a 1550 nm linearly polarized beam was firstly boosted to 30 mW in a custom-built erbium doped fiber amplifier (EDFA) and its emission was split into the probe and reference beams using a fiber coupler (10 dB). The probe beam was outcoupled from the fiber into free space using a fiber collimator (FC), and coupled into the ARHCF-based GAC with the aid of a NBK-7 plano-convex lens (f=100 mm). The near-IR light exiting the far side of the GAC is separated from the mid-IR beam by a germanium wedge, reflected by a silver mirror, focused and collected via a FC, reaching a 1.15 mW optical power at the photodiode. The reference beam undergoes an Ω=40 MHz frequency downshift in an acoustic-optical modulator (AOM) and is combined with the probe beam in a second fiber coupler (3 dB) and delivered to an Indium Gallium Arsenide (InGaAs) photodiode to detect the heterodyne beatnote signal. An additional radio frequency (RF) amplifier (PE15A1012) was used to boost the signal by 20 dB. If we consider that the NO molecules are experiencing a periodical excitation by the pump laser in the GAC, a RI modulation will occur, which will be observable along the probe arm of the MZI as a phase modulation of the probe light. The intensity of the phase change Δθ can be expressed as follows [19]:
where C is the gas concentration, α(λpump) is the absorption coefficient at the pump wavelength, L is the length of the MZI arm where the RI modulation occurs, Ppump is the pump power of the laser inducing the RI modulation. The k parameter depends on the gas parameters, modulation frequency of the pump light and the guidance parameters of the fiber used as the GAC (i.e. mode field diameter for pump and probe, beam overlap, etc.). The phase modulation amplitude is linearly dependent on the gas molecules concentration, the pump power and the length of the gas-laser interaction length, thus can be conveniently used to determine the concentration of the target gas. In the case of a heterodyne MZI, the phase variation and its amplitude can be detected by a simple frequency demodulation of the heterodyne MZI signal at the beat-note frequency Ω. In the presented experiment an RF spectrum analyzer (Rohde Schwarz, FSV3007) was used to demodulate the beatnote signal at Ω=40 MHz. The demodulated signal was digitized and processed in a custom LabView-based application, which calculates the FFT of the acquired spectroscopic signal, enabling in-depth analysis of the amplitude of the harmonics of the sinewave modulation frequency - f0, especially focusing on its second harmonic - 2·f0. In the experiments a counter-propagating pump and probe configuration was chosen as it minimizes the number of bulk optics based components in the sensor. All components used to setup the probe section of the sensor were polarization maintaining. Both MZI arms were cut to approximately the same length (±3 mm). The proposed configuration utilizes the most significant advantage of the PTS – the separation of the pump and the probe part of the sensor. The results clearly show that with proper ARHCF selection, the pump source can be chosen to target the strongest absorption profiles of the analyte, while the PTS signal retrieval can be realized based on inexpensive, reliable and widely available telecom fiber technology-based probe.3. Experimental results
3.1 Optimization of the WMS parameters
During the experiments a certified mixture of 300 parts-per-million by volume (ppmv) NO in nitrogen (N2) was used, which was diluted using a commercial gas mixer system (Environics, 4020), if needed. The effect of the RI modulation induced by the sinewave-modulated pump laser can be clearly observed in the electric signal delivered by the photodiode monitoring the output of the MZI. RF spectra of the signals registered for N2 and 300 ppmv of NO, flowing through the GAC are presented in Fig. 2.
Fig. 2. Effect of the photothermal RI modulation on the MZI signal. RF spectrum of the detector signal registered for N2 (a) and 300 ppmv NO (b) flowing through the GAC. Clear sidebands corresponding to the 2.f0 modulation can be observed around the carrier frequency Ω = 40 MHz when the NO gas molecules are introduced into the GAC and are experiencing photothermal excitation.
For this demonstration, the pump laser was tuned to the center of the NO transition at 1900.09 cm-1. The registered signals clearly show the difference in the RF spectrum for N2 and NO flowing through the GAC, respectively. When the fiber was filled with NO molecules, evenly spaced modulation sidebands separated by 12 kHz from the carrier frequency (Ω=40 MHz) were observed in the RF spectrum (QCL pump sinewave modulation frequency was set to 6 kHz). To achieve the best signal-to-noise ratio (SNR) of the constructed sensor the optimal parameters had to be determined. Firstly, I focused on the optimal pump modulation frequency – f0. The FG used in the experiment to generate the sinusoidal modulation and the mHz sawtooth ramp was capable of producing a frequency sweep. This functionality was used to scan the QCL modulation frequency - f0, between 100 Hz and 50 kHz, which was used to determine the optimal value of this parameter. During the measurement, the QCL laser was tuned to the center of the NO transition at 1900.09 cm-1, the MZI signal was demodulated and an FFT was calculated via the LabView application, which monitored the amplitude at the second harmonic (2.f0) of the f0 modulation frequency. Figure 3 shows the results of the measurement.
Fig. 3. Pump laser sinewave modulation frequency optimization. (a) 2f signal amplitude plotted as a function of the sinewave modulation of the pump laser. (b) Full 2f scans of the NO doublet collected for different values of the pump modulation frequency. Measurement parameters are listed in graph (a).
Based on the gathered experimental results, the optimum sinewave modulation frequency for the presented sensor configuration was 6 kHz. This value was used in each of the following experiments. The characteristic plotted in Fig. 3(a) deviates from an exponential fit for frequencies below 5 kHz due to an digital highpass filter used in the demodulator, which ensures proper noise rejection. The second parameter requiring precise fine-tuning in the WMS detection technique is the sinewave modulation depth. The FG was set to sweep the amplitude of the generated sinewave modulation between 3.5 GHz and 14.5 GHz. The measurement technique was similar to the abovementioned. The results of the measurements are depicted in Fig. 4.
Fig. 4. (a) Sensor response for varying pump sinewave modulation depth. (b) full 2f WMS scans registered for three values of the modulation depth. Measurement parameters are listed in graph (a).
Based on the performed measurements, the optimal modulation depth at which the 2f signal amplitude reaches the maximum was 8 GHz. This value was used in the following experiments. Figure 4(b) shows full 2f scans across the absorption line doublet of NO for several setpoints of the modulation depth. Note that for a modulation depth of 12.5 GHz the signals from the individual NO absorption lines begin to merge severely.
3.2 Experimental results of NO detection at 5.26 µm
The performance of the sensor was evaluated based on the measurement of a certified mixture of NO in N2 with a concentration equal to 300 ppmv flowing through the sensor at 800 Torr. During the measurements the sinewave modulation was set to f0 = 6 kHz with a modulation depth of 8 GHz, as evaluated in subsection 3.1. The maximum 2f WMS signal registered for a full scan is shown in Fig. 5(a).
Fig. 5. (a) – WMS signal registered with a full scan across the NO doublet transition for optimized parameters (listed on the graph). (b) – HITRAN-based simulation of absorption spectra for ambient air and 300 ppmv of NO diluted in N2 simulated for 800 Torr and a 25 cm long pathlength [25]. The targeted strong transition is highlighted with an arrow.
With the optimized WMS parameters the maximum 2f signal amplitude reached 43090 units for a pump power of 82 mW. The targeted absorption line (Fig. 5(b)) was interference-free from molecules normally present in air e.g. water vapor or carbon dioxide, thus the registered signal shows no background disturbances.
The linearity of the constructed sensor was estimated by diluting the certified 300 ppmv NO in N2 mixture using a commercial gas mixer (Environics, 4020). During the experiment, the pump laser was tuned to the center of the NO transition located at 1900.09 cm-1 and the 2f WMS signal amplitude was registered and averaged for 60 seconds for each dilution point. The results are plotted in Fig. 6.
Fig. 6. Maximum 2f WMS amplitude plotted as a function of NO concentration flowing through the gas cell at 800 Torr. Red solid line represents a linear fit of the measured values.
The performed measurement confirmed that the maximum 2f signal increases approximately linearly with the measured gas concentration. A linear fit of the registered values yielded an R-square (R2) value of 0.9989, which is comparable with previously published experimental work on similar PTS-based gas sensors [7,19,20]. A detailed comparison of the sensor readout for pure N2 and 2 ppmv of NO flowing through the sensor (at 800 Torr pressure), respectively, is depicted in Fig. 7. During the measurement the pump laser was tuned to the center of the selected NO transition, as in previous experiments.
Fig. 7. (a) – maximum 2f WMS signal registered for pure N2 and 2 ppmv NO flowing through the sensor. At t = 0 the GAC began to be filled with 2 ppmv of NO. (b) – demodulated signal at Ω=40 MHz for pure N2 (left graph) and 2 ppmv NO (right graph), respectively.
No background signal is observed when the GAC is filled with pure N2 and no DC offset is present in the measurement, apart from random noise. At t = 0 the GAC began to be filled with 2 ppmv of NO and the registered 2f signal amplitude promptly increased to a value of ∼280 a.u. Although the 2 ppmv of NO has an absorption coefficient of only 3.16·10−5 cm-1, and the GAC was only 25 cm long, the registered 2f signal was clearly distinguishable from the N2 background noise. A clear difference is also visible when the raw output of the analog demodulator is examined for both cases. For N2 flowing through the GAC only the f0 frequency is present in the demodulated signal however, when 2 ppmv of NO is flowing through the sensor, the second harmonic of the 6 kHz sinewave modulation of the pump laser is clearly evident and reaches a frequency deviation equal to 2.58 Hz. The residual f0 modulation seen on the left panel in Fig. 7(b) is observable in the demodulated signal due to the absorption of the mid-IR pump beam in the silica SMF28 fiber (fiber collimator and fiber delivering the probe beam to the interferometer). Due to the approximately linear absorption spectrum of silica at the pump wavelength (5.25 µm), the modulation does not experience a nonlinearity if no gas is present in the absorption path, which would result in generating harmonics of the modulation, having a negative impact on the detection limit of the constructed sensor. The process of extracting the photothermal modulation in the constructed sensor is as follows. Because the MZI has an acousto-optical frequency shifter in the reference arm, a strong beatnote frequency 40 MHz is observable (see Fig. 2). If no NO gas molecules are present in the sensor the RI on the pathlength of the interferometer arm will not be modulated, thus the carrier beatnote will not move (apart from the miniscule residual modulation at f0). However, if the 6 kHz modulated pump beam excites NO molecules in the interferometer arm the 40 MHz beatnote will experience a frequency shift proportional to the intensity of the RI modulation. This shift can be easily detected by demodulating the signal at the carrier frequency (Ω=40MHz). Here this is realized by an analog demodulator inbuilt into the RF spectrum analyzer. A sample of the demodulated signal is presented in Fig. 7(b). The demodulated signal is digitized and processed by a custom LabView application. The application calculates the FFT of the demodulated signal and plots the amplitude of the f0 and 2·f0 components as a function of time. A similar effect could be obtained by monitoring the intensity of the sidebands observable in the raw RF spectrum presented in Fig. 2, however, it provides inferior response time and sensitivity, when compared with the one used in this experiment. During the performance evaluation I have also examined the 3 hour stability of the sensor and the gas exchange time in the GAC, which are crucial from the application point of view. Results of the experiments are depicted in Fig. 8.
Fig. 8. a) – Closeup of several full 2f WMS scans registered for 300 ppmv of NO flowing through the GAC during a 3 hour measurement. The inset shows the scans registered during a 3 hour-long measurement. b) – 2f signal amplitude measured as a function of time during gas loading into the GAC.
To examine the long-term stability of the sensor, I have monitored the amplitude of the 2f WMS scans during 3 h of continuous operation with 300 ppmv of NO flowing through the sensor (180 full scans were recorded). The highest discrepancy of the maximum 2f amplitude during this measurement reached 3.1%, which is comparable with similar configurations based on the PTS approach [19]. Due to the fact, that the PTS signal is encoded in frequency, not in amplitude, I believe, that the main source of the sensor instability recorded in Fig. 8(a) is connected with the fluctuation of the gas pressure and flow, which can be easily mitigated by using an active pressure-regulated gas delivery system. To measure the gas exchange time, the GAC was firstly flushed with pure N2, the pump laser was tuned to the center of the target NO transition at ∼1900.09 cm-1 and the 2f amplitude of the signal was measured for 50 s. Next, the valves were cycled and the sensor was loaded with 300 ppmv of NO in N2 using an overpressure of 40 Torr. At t=0 s the registered 2f signal began to increase and reached 90% of the maximum amplitude after 65 s. The total volume of the GAC is equal to 8.12 µL, which gives a flow of ∼7.5 × 10−3 sccm. The obtained gas exchange time in my sensor is longer than in the ARHCF-based gas sensors relying on the absorption-based gas detection techniques [3], where the pressures used to force the flow of gas have less influence on the noise of the sensor, thus can be higher. Nevertheless, when compared to sensors based on the PT effects with e.g. interferometric signal readout, the filling times are comparable [9].
3.3 Sensor performance evaluation
In Fig. 9, the performance of the MZI PTS NO sensor was evaluated based on calculating the Allan deviation from a 25 minute-long measurement of noise of the sensor with N2 flowing through the GAC (depicted in Fig. 9(a)).
Fig. 9. (a) Time trace of the sensor baseline noise measured for 25 minutes for the pump laser turned ON and tuned to the peak of the NO absorption with pure N2 flowing through the GAC. (b) Allan deviation plot calculated based on the sensor baseline noise from Fig. 9(a) expressed in ppbv units. Dashed red line represents the white noise trend (1/√t). Minimum detection limits of the sensor for 1 s and 100 s are highlighted in the graph.
The dashed red line in Fig. 9(b) represents the 1/√t noise trend, in which the white noise dominates as the source of instability in the sensor. The calculated Allan deviation plot closely follows the 1/√t trend without deviation for an averaging time up to 400 s, which clearly shows the benefits of using the purely frequency-based (heterodyne) PTS signal readout used in my sensor configuration, when compared to sensors relying on amplitude monitoring [6,7,9]. Y-axis of the plot was expressed in ppbv units based on the maximum 2f signal amplitude registered for a certified 300 ppmv mixture of NO in N2 and optimized WMS parameters, depicted in Fig. 5(a). The sensor reached an MDL equal to 18 ppbv for 1 s and 1.78 ppbv for 100 s integration time, respectively.
4. Discussion and conclusions
Photothermal gas sensing techniques have proven to provide exceptionally low detection limits and high selectivity. Such sensors use a pump laser, which absorbed by the target gas causes a localized heating of the sample, modulating the RI. This RI modulation can be observed as a phase modulation of the probe beam transmitted through the sample. Due to miniscule phase changes, the readout of the PTS signal is usually accomplished using interferometric techniques. The most significant advantage of PTS is the possibility to separate the gas pump laser from the probe laser. This enables using high power QCLs for inducing the RI modulation, while proper design allows incorporating robust and inexpensive near-IR fiber components to setup the probe part of the sensor. In this sensor, I have taken advantage of this unique feature and designed a heterodyne MZI based readout of the induced RI modulation (phase modulation). This approach is purely frequency-based – the spectroscopic signal is encoded in frequency, not in the amplitude of the MZI signal, therefore mitigating the numerous drawbacks identified by researchers using homodyne MZI or FPI in combination with PTS and HCFs [6–8,15,19]. Homodyne detection is prone to acoustic and mechanical noise sources and long-term drifts of the coupling optics severely influence the performance of the sensor. Moreover, the interferometers have to be actively stabilized precisely in the quadrature point, to achieve a linear translation of the RI modulation to the photodiode amplitude signal output. This task is not trivial and requires using piezoceramic fiber stretchers with limited frequency response, or using wavelength locking of the probe laser to the quadrature point. Furthermore, PTS detection of gases with very high concentrations could result in a phase modulation exceeding the linear part of the detection slope, which requires additional signal processing methods to correctly decode the gas concentration values. In the case of heterodyne detection, the PTS signal is encoded in frequency, not in amplitude of the beam reaching the detector, thus random amplitude variations, or long-term drift will influence the noise level significantly less. Moreover, the linearity of the sensor response is limited only by the bandwidth of the demodulator used to convert the frequency deviation to a signal interpretable e.g. by a traditional lock-in amplifier. The modification of the signal readout from an amplitude method (e.g. FPI, homodyne MZI) to the frequency-based resulted in reaching lower detection limits, compared to previous mid-IR HCF-based gas sensors. The sensor presented in this paper reached a noise equivalent absorption equal to 2.86×10−7 cm-1 and 2.81×10−8 cm-1, for 1 s and 100 s integration times, respectively. The performance of the proposed sensor is at least an order of magnitude better than reported in previous publications and to the best of my knowledge is currently the lowest sensitivity obtained in the sensor utilizing a mid-IR ARHCF as a GAC [7,6,14,15,19,21]. Moreover, because the Allan deviation plot closely follows the 1/√t trend, thus a longer averaging time can be used to achieve lower detection limits if required and applicable to the application. The performance of the selected HCF-based mid-IR gas sensors was summed up in Table 1.
Table 1. Performance of selected gas sensors based on mid-IR HCF-based GACs.
I have used a commercial apparatus (RF spectrum analyzer) for analog demodulation of the heterodyne signal in my setup, which subsequently required calculating the FFT with the aid of a custom LabView application to readout the spectroscopic signal. This approach was far from cost-effective and field-deployable and was the main limiting factor for achieving higher signal levels. However, a similar result could be obtained by using integrated circuits for this purpose (e.g. AD 630, using frequency downmixing of the beat note to match the chip bandwidth) combined with a portable FPGA-based lock-in amplifier (e.g. Zurich Instruments, or other self-programable platforms). The ultimate phase modulation signal retrieval could be obtained by employing a custom developed IQ demodulator, which are known to provide exceptional sensitivity in heterodyne interferometry [22]. I believe, that further development of novel types of mid-IR hollow core fibers and gas sensors based on them will lead to obtaining detection limits comparable with complex state-of-the-art detectors using cavity enhanced techniques [23] or hundred-meter long multipass cells [24].
Funding
Narodowe Centrum Nauki (2019/01/Y/ST7/00088, 2019/35/D/ST7/04436).
Acknowledgments
K. Krzempek thanks Walter Belardi for providing the ARHCF, Pawel Kozioł for manufacturing the microchannels in the ARHCF and Piotr Jaworski for proofreading the paper.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data that supports the findings of this study is available from the corresponding author upon reasonable request.
References
1. W. Belardi and P. J. Sazio, “Borosilicate Based Hollow-Core Optical Fibers,” Fibers 7(8), 73 (2019). [CrossRef]
2. W. A. Challener, A. M. Kasten, F. Yu, G. Puc, and B. J. Mangan, “Dynamics of Trace Methane Diffusion/Flow Into Hollow Core Fiber Using Laser Absorption Spectroscopy,” IEEE Sens. J. 21(5), 6287–6292 (2021). [CrossRef]
3. P. Jaworski, K. Krzempek, G. Dudzik, P. J. Sazio, and W. Belardi, “Nitrous oxide detection at 5.26 µm with a compound glass antiresonant hollow-core optical fiber,” Opt. Lett. 45(6), 1326–1329 (2020). [CrossRef]
4. M. Nikodem, K. Krzempek, G. Dudzik, and K. Abramski, “Hollow core fiber-assisted absorption spectroscopy of methane at 3.4 µm,” Opt. Express 26(17), 21843–21848 (2018). [CrossRef]
5. M. Nikodem, G. Gomółka, M. Klimczak, D. Pysz, R. Buczyński, and R. Buczyński, “Demonstration of mid-infrared gas sensing using an anti-resonant hollow core fiber and a quantum cascade laser,” Opt. Express 27(25), 36350–36357 (2019). [CrossRef]
6. Z. Li, Z. Wang, F. Yang, W. Jin, and W. Ren, “Mid-infrared fiber-optic photothermal interferometry,” Opt. Lett. 42(18), 3718–3721 (2017). [CrossRef]
7. F. Yang, Y. Tan, W. Jin, Y. Lin, Y. Qi, and H. L. Ho, “Hollow-core fiber Fabry–Perot photothermal gas sensor,” Opt. Lett. 41(13), 3025–3028 (2016). [CrossRef]
8. C. Yao, S. Gao, Y. Wang, P. Wang, W. Jin, and W. Ren, “MIR-Pump NIR-Probe Fiber-Optic Photothermal Spectroscopy With Background-Free First Harmonic Detection,” IEEE Sens. J. 20(21), 12709–12715 (2020). [CrossRef]
9. H. Bao, H. Bao, Y. Hong, Y. Hong, W. Jin, W. Jin, H. L. Ho, H. L. Ho, C. Wang, C. Wang, S. Gao, Y. Wang, and P. Wang, “Modeling and performance evaluation of in-line Fabry-Perot photothermal gas sensors with hollow-core optical fibers,” Opt. Express 28(4), 5423–5435 (2020). [CrossRef]
10. S. E. Bialkowski, N. G. C. Astrath, and M. A. Proskurnin, Photothermal Spectroscopy Methods (John Wiley & Sons, 2019).
11. C. C. Davis and S. J. Petuchowski, “Phase fluctuation optical heterodyne spectroscopy of gases,” Appl. Opt. 20(14), 2539–2554 (1981). [CrossRef]
12. K. Krzempek, A. Hudzikowski, A. Głuszek, G. Dudzik, K. Abramski, G. Wysocki, and M. Nikodem, “Multi-pass cell-assisted photoacoustic/photothermal spectroscopy of gases using quantum cascade laser excitation and heterodyne interferometric signal detection,” Appl. Phys. B 124(5), 74 (2018). [CrossRef]
13. M. A. Owens, C. C. Davis, and R. R. Dickerson, “A Photothermal Interferometer for Gas-Phase Ammonia Detection,” Anal. Chem. 71(7), 1391–1399 (1999). [CrossRef]
14. Y. Lin, W. Jin, F. Yang, J. Ma, C. Wang, H. L. Ho, and Y. Liu, “Pulsed photothermal interferometry for spectroscopic gas detection with hollow-core optical fibre,” Sci. Rep. 6(1), 39410 (2016). [CrossRef]
15. C. Yao, Q. Wang, Y. Lin, W. Jin, L. Xiao, S. Gao, Y. Wang, P. Wang, and W. Ren, “Photothermal CO detection in a hollow-core negative curvature fiber,” Opt. Lett. 44(16), 4048–4051 (2019). [CrossRef]
16. P. Zhao, Y. Zhao, H. Bao, H. L. Ho, W. Jin, S. Fan, S. Gao, Y. Wang, and P. Wang, “Mode-phase-difference photothermal spectroscopy for gas detection with an anti-resonant hollow-core optical fiber,” Nat Commun 11(1), 847 (2020). [CrossRef]
17. B. Fang, N. Yang, C. Wang, W. Zhao, X. Xu, Y. Zhang, and W. Zhang, “Detection of nitric oxide with Faraday rotation spectroscopy at 5.33 µm,” Chin. J. Chem. Phys. 33(1), 37–42 (2020). [CrossRef]
18. K. Krzempek, G. Dudzik, K. Abramski, G. Wysocki, P. Jaworski, and M. Nikodem, “Heterodyne interferometric signal retrieval in photoacoustic spectroscopy,” Opt. Express 26(2), 1125–1132 (2018). [CrossRef]
19. F. Chen, F. Chen, S. Jiang, W. Jin, W. Jin, H. Bao, H. Bao, H. L. Ho, H. L. Ho, C. Wang, C. Wang, and S. Gao, “Ethane detection with mid-infrared hollow-core fiber photothermal spectroscopy,” Opt. Express 28(25), 38115–38126 (2020). [CrossRef]
20. J. P. Waclawek, H. Moser, and B. Lendl, “Balanced-detection interferometric cavity-assisted photothermal spectroscopy employing an all-fiber-coupled probe laser configuration,” Opt. Express 29(5), 7794–7808 (2021). [CrossRef]
21. P. Jaworski, P. Kozioł, K. Krzempek, D. Wu, F. Yu, P. Bojęś, G. Dudzik, M. Liao, K. Abramski, and J. Knight, “Antiresonant Hollow-Core Fiber-Based Dual Gas Sensor for Detection of Methane and Carbon Dioxide in the Near- and Mid-Infrared Regions,” Sensors 20(14), 3813 (2020). [CrossRef]
22. S. Yoon, Y. Park, and K. Cho, “A new balanced-path heterodyne I/Q-interferometer scheme for low environmental noise, high sensitivity phase measurements for both reflection and transmission geometry,” Opt. Express 21(18), 20722–20729 (2013). [CrossRef]
23. L. Richard, D. Romanini, and I. Ventrillard, “Nitric Oxide Analysis Down to ppt Levels by Optical-Feedback Cavity-Enhanced Absorption Spectroscopy,” Sensors 18(7), 1997 (2018). [CrossRef]
24. D. D. Nelson, J. H. Shorter, J. B. McManus, and M. S. Zahniser, “Sub-part-per-billion detection of nitric oxide in air using a thermoelectrically cooled mid-infrared quantum cascade laser spectrometer,” Applied Physics B: Lasers and Optics 75(2-3), 343–350 (2002). [CrossRef]
25. L. S. Rothman, I. E. Gordon, A. Barbe, D. C. Benner, P. F. Bernath, M. Birk, V. Boudon, L. R. Brown, A. Campargue, J.-P. Champion, K. Chance, L. H. Coudert, V. Dana, V. M. Devi, S. Fally, J.-M. Flaud, R. R. Gamache, A. Goldman, D. Jacquemart, I. Kleiner, N. Lacome, W. J. Lafferty, J.-Y. Mandin, S. T. Massie, S. N. Mikhailenko, C. E. Miller, N. Moazzen-Ahmadi, O. V. Naumenko, A. V. Nikitin, J. Orphal, V. I. Perevalov, A. Perrin, A. Predoi-Cross, C. P. Rinsland, M. Rotger, M. Šimečková, M. A. H. Smith, K. Sung, S. A. Tashkun, J. Tennyson, R. A. Toth, A. C. Vandaele, and J. Vander Auwera, “The HITRAN 2008 molecular spectroscopic database,” Journal of Quantitative Spectroscopy and Radiative Transfer 110(9-10), 533–572 (2009). [CrossRef]
26. K. Krzempek, K. Abramski, and M. Nikodem, “Kagome Hollow Core Fiber-Based Mid-Infrared Dispersion Spectroscopy of Methane at Sub-ppm Levels,” Sensors 19(15), 3352 (2019). [CrossRef]
27. C. Yao, S. Gao, Y. Wang, P. Wang, W. Jin, and W. Ren, “Silica Hollow-Core Negative Curvature Fibers Enable Ultrasensitive Mid-Infrared Absorption Spectroscopy,” J. Lightwave Technol. 38(7), 2067–2072 (2020). [CrossRef]
