Open Access
Add to BibSonomy
Abstract
An innovative fiber-enhanced Raman gas sensing system with a hollow-core anti-resonant fiber is introduced. Two iris diaphragms are implemented for spatial filtering, and a reflecting mirror is attached to one fiber end that provides a highly improved Raman signal enhancement over 2.9 times than the typical bare fiber system. The analytical performance for multigas compositions is thoroughly demonstrated by recording the Raman spectra of carbon dioxide (CO2), oxygen (O2), nitrogen (N2), hydrogen (H2), and sulfur dioxide (SO2) with limits of detection down to low-ppm levels as well as a long-term instability < 1.05%. The excellent linear relationship between Raman signal intensity (peak height) and gas concentrations indicates a promising potential for accurate quantification.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Highly sensitive gas sensing has essential applications in fields of atmospheric monitoring, industrial process control, and biomedical analysis. Spectroscopic techniques outperform the traditional sensing methods, such as gas chromatography (GC), semiconductor and electrochemical sensors, with the superiority of non-invasiveness, no sample-consumption, high selectivity, and potential of in-situ real-time measurement. Furthermore, the advanced approaches based on infrared absorption spectroscopy (IAS) [1–3], photoacoustic spectroscopy (PAS) [4–11] / photothermal spectroscopy (PTS) [12–14], and noise-immune cavity-enhanced optical heterodyne molecular spectroscopy (NICE-OHMS) [15–17] can even provide ultra-sensitivity for some gases, pushing the limits of detection (LODs) down to extremely low-ppb levels. However, these approaches fail to work on the high-profile homonuclear diatomic gases (e.g., H2, O2, N2, Cl2) and indispensably require several laser sources with different wavelengths as well as the wide-band detectors for multigas analysis.
Raman spectroscopy is an emerging versatile spectroscopic technique that provides the distinctive capability of identifying and quantifying almost all components of the gaseous mixtures within a wide concentration range using one single-wavelength laser source, including the isotopic molecules. However, this technique is not yet completely established due to the insufficient LODs for practical applications attributed to the low probability of the Raman scattering process. Cavity-enhanced Raman spectroscopy (CERS) using high-finesse optical resonant cavities [18–22] or multipass cells [23–25] is an elaborate enhancement method that can provide application-required low-ppm and even sub-ppm sensitivity, but it is usually not well-suited for field-based gas sensing with a limited sample. Fiber-enhanced Raman spectroscopy (FERS) is a promising alternative candidate that enables enhanced light-analyte interaction on extended long pathlength and integrates the signal collection angles over the whole fiber length using hollow-core optical fibers (HCFs), usually including the metal-coated capillaries (MCCs) [26–28] and micro-structured optical fibers (MSFs) [29–36].
To the best of our knowledge, William et al. [26,27] firstly described a silver-coated MCC system for Raman measurement and achieved a 30-fold signal enhancement for ambient N2. Nevertheless, MCC suffers from high bend loss, necessitating it being configured straight on the optical axis with minimal bending, and the achievable LODs (typically in hundreds of ppm) are not satisfactory due to the high transmission loss and strong fluorescence induced by the interaction of the laser-glass components in the setup. Fortunately, these problems can be avoided by employing MSFs, such as hollow-core photonic bandgap fiber (HC-PBF) and hollow-core anti-resonant fiber (HC-ARF). Buric et al. [29,30] established the basic framework of Raman gas sensing system with HC-PBFs, in which a complex spatial filter system consisting of two aspheric lenses and a pinhole was required to eliminate the background silica noise for a better detection sensitivity, and the LODs for investigated gases were push down to ppm-level. Hanf. et al. [31,32] further developed the HC-PBF system and applied it in environmental gas sensing as well as human breath analysis with highly improved analytical sensitivity down to sub-ppm LODs. Subsequently, Knebl et al. [34–36], coming from the same research group, firstly utilized the newly fabricated HC-ARF for Raman gas sensing and demonstrated its applications in the environmental sciences. However, the small core diameters (∼ 5 to ∼30 μm) of MSFs usually lead to a long gas filling time (several hours for meters-long MSF) under the free diffusion condition [37,38]. Vacuum/pressure-assisted filling is an effective method for speeding up the gas delivery into the fiber, and a response time of several seconds has been achieved in the laboratory [33,34]. Nevertheless, the real-time gas sensing cases that are absent of a pressure differential in practice may only depend on the diffusion-based gas loading techniques (e.g. gas sensing in the atmosphere environment, or in the gas-insulated electrical equipment where the pressure is typically up to 5 bar). Introduction of side-microchannels along MSFs is an alternative approach [39–41], which can work as the additional channels for gas delivery into MSFs under the free diffusion conditions, equivalent to shorten the fiber length, and thus speeding up the gas filling process. The validity of these microchannels has been demonstrated by Hoo et al. [39], where a diffusion-limited response time was reduced to ∼ 3 s for a 7-cm HC-PBF with seven periodic side-microchannels (separated by 1 cm) compared to ∼ 55 s for the 7-cm HC-PBF with both ends open and without microchannels.
In this paper, we introduce an innovative fiber-enhanced Raman gas sensing system using a customized nodeless HC-ARF, where two iris diaphragms with optimized diameters are implemented for high-efficiency spatial filtering that improves the signal to noise (SNR) up to 5-fold. A reflecting mirror attached to the rear fiber end is employed for effective signal enhancement, which provides a gain factor over 2.9 than the typical bare FERS system. Several microchannels are drilled along the fiber side, proving the FERS technology with microchannel-modified HC-ARF is feasible. The performance of the system is demonstrated by obtaining the Raman spectra of ambient air in the laboratory, H2, and SO2, which shows that the FERS technique has promising potential in trace gas sensing field with high selectivity and sensitivity, as well as the feasibility for accurate quantification.
2. Experimental details
2.1 Design of the HC-ARF gas cell
The fiber-enhanced Raman gas cell is constructed with a 0.27-m-long HC-ARF drilled with 6 microchannels along the side using a 1030 nm femtosecond laser (Changzhou Keedy Laser Technology Co., Ltd.), as shown in Fig. 1(a), for achieving fast gas filling under free diffusion and providing potential applications in the real-time gas sensing cases that are absent of a pressure differential. The separations between these microchannels are 4 cm, and the core diameters range from ∼1 to ∼2 μm near the hollow fiber core, as shown in Fig. 1(b). Both ends of the HC-ARF are carefully cleaved, and the output end is coupled to a reflecting mirror (MHCF), that is glued on the end facet of a ceramic ferrule (CF1) and provides an average reflectivity of approximately 99% for the light ranging from 400 to 750 nm, for high-efficiency Raman signal enhancement, which has been demonstrated in [42,43], as shown in Fig. 1(c).
Fig. 1. Structure of the 0.27-m-long HC-ARF gas cell and customized fiber adapter assemblies. (a) Picture of the HC-ARF gas cell drilled with 6 microchannels along the side, the locations of which are indicated with red dotted circles. The rear end of the HC-ARF is coupled to a reflecting mirror and assembled by the ceramic ferrules. In the experiment, the HC-ARF is fixed onto a 0.22-m-long, 2-mm-wide thin aluminum sheet by the UV-curable adhesive and then placed inside a 0.15-m-long, 4-mm-diameter metal tubular gas chamber with both ends connected to the customized fiber adapters. (b) Scanning electron microscopy (SEM) image of the fiber cross-section with one typical side-drilled microchannel, the core diameter of which is about 1.02 μm near the hollow core. (c) Structure of the fiber-mirror coupling assembly, consisting of two ceramic ferrules for fixing and a coaxial sleeve for alignment. (d) Structure of the customized fiber adapters, mainly consisting of the fixtures, adapter, V-groove holder, optical window, seal rings, and other mechanical components.
The fabrication procedure of the HC-ARF gas cell is described as follows: the 0.27-m-long HC-ARF with microchannels is kept straight and carefully fixed onto a 0.22-m-long, 2-mm-wide thin aluminum sheet by the UV-curable adhesive and then placed inside a 0.15-m-long, 4-mm-diameter metal tubular gas chamber with both ends connected to the customized fiber adapters (FA1, FA2), as shown in Fig. 1(d). The front end of the HC-ARF is glued into a standard V-groove fiber holder (Thorlabs: HFV001) with blu-tack and then assembled in FA1; the rear end is inserted and fixed to another ceramic ferrule (CF2) with the inner diameter of 350 μm and outer diameter of 2.5 mm by the UV-curable adhesive, the assembly is then inserted into a coaxial sleeve with an inner diameter of 2.5 mm and glued into a customized metal V-groove holder with blu-tack and installed in FA2. The two ceramic ferrules (CF1 & CF2) are joined tightly and auto-aligned by the coaxial sleeve. A small gap is left between the end facet of the HC-ARF and the reflecting mirror, which serves as the channel for gas filling and flushing just as the 0.5-mm-wide slit in the coaxial sleeve does. At the beginning of the experiment, the ceramic ferrule (CF1) glued with a reflecting mirror is not inserted into the coaxial sleeve, for we have to monitor the transmitted laser power at the rear fiber end. Once a high-efficiency laser-fiber coupling (> 85%) is achieved, the ceramic ferrule (CF1) would be inserted, and thus the reflecting mirror is attached closely to the rear fiber end. In experiments, FA1 is connected with a gas cylinder for gas filling, and FA2 is connected to a vacuum pump for exhausting. Since FA1 and FA2 are directly connected via the metal tube, the gases can be filled into or exhausted from the HC-ARF from two ends and there is no almost pressure difference between them.
2.2 Optical setup for FERS system
The optical design for the compact and robust HC-ARF-based Raman gas sensing system is schematically illustrated in Fig. 2. A continuous-wave diode-pumped solid-state laser (CW-DPSS) (with λL = 532 nm, IL = 1.5 W, and M2 < 1.1) from Cobolt is applied as the light source. The output linear polarization wave (with a polarization ratio >100:1) passes through a half-wave plate (HWP, Thorlabs: WPH05M-532) to adjust its polarization direction and then is guided by a polarization beam splitter (PBS, Thorlabs: WPBS254-VIS), thus controlling the laser output power for eye-safety in optical alignment. The s-polarized laser beam collimated via two wide-band reflectors (M1, M2) after the PBS is reflected into an achromatic lens (AL1) with a selected focal length f1 = 60 mm, at the focal spot location of which the laser beam size matches the mode field diameter (MFD) of the HC-ARF, by a dichroic beam splitter (DBS, Semrock: LPD02-532RU-25) with an edge at 537 nm, providing a reflectivity of 98% for the s-polarized laser radiation incident at 45° and transmissivity of 93% for the red-shifted Raman signal. The focused spot of the laser beam is spatially adjusted via a 3-axis nanopositioning stage (3A-S, Sigma: TSD-652-M6 for XY and TSD-653-M6 for Z) and then precisely coupled into a 0.27-m-long HC-ARF with a broad transmission bandwidth from 440 to 1200 nm [44]. The core diameter of the HC-ARF is about 26 μm that reduces the difficulty for laser coupling and allows for a relatively robust measurement compared with the previously used hollow-core photonic bandgap fiber (HC-PBF, core diameter of approximately 7 μm) [29–33]. A power meter (PM) monitors the transmitted laser power for calculating the laser-fiber coupling efficiency and providing guidance for the optical alignments.
Fig. 2. Schematic sketch of the fiber-enhanced Raman gas sensing system. The laser beam (central wavelength at 532 nm) is guided by a half-wave plate (HWP), a polarization beam splitter (PBS), two reflectors (M1, M2) for collimation, a dichroic beam splitter (DBS), another reflector (M3) for redirection and an achromatic lens (AL1) into the hollow-core anti-resonant fiber (HC-ARF, core diameter of 26 μm). The backscattered Raman signal is collected by the original achromatic lens (AL1), passing through two iris diaphragms (ID1, ID2) for spatial filtering, the dichroic beam splitter and two edge filters (EF1, EF2) for suppressing the Rayleigh scattered light, and coupled into the slit of the spectrometer (SPEC) by the other achromatic lens (AL2). The gases are introduced into / emptied from the hollow fiber core through two custom-made fiber adapters (FA1, FA2) at both fiber ends. A power meter (PM) monitors the transmitted laser power for calculating the coupling efficiency.
The backscattered Raman signal is collected and collimated via the previous achromatic lens (AL1), then the red-shifted light beam successively passes through the dichroic beam splitter (DBS) and two edge filters (EF, Semrock: LP03-532RE-25), where the Rayleigh scattered signal is almost completely suppressed, and eventually is focused via the other achromatic lens (AL2, focal length f2 = 75 mm) into the slit of a spectrometer (SPEC) from Andor (Model: Shamrock-500i) in which the beam is diffracted and guided to a high-performance charge-coupled device (CCD) from Andor (Model: DU416A-LDC-DD) to be detected. The spectrometer allows a wide spectral observation range extending from 100 to 4500 cm-1, covering all relevant molecules of interest for gas-phase Raman spectroscopy. A grating with 1200 lines per millimeter (blaze wavelength at 750 nm), which provides a high diffraction efficiency and spectral resolution (0.06 nm), and a slit size of 70 μm are selected for all Raman measurements, if not otherwise specified.
What is worth mentioning is that two iris diaphragms (ID1, ID2, Thorlabs: SM1D12D) with the optimum core diameters are inserted between DBS and M3, M4 and EF1, respectively, to filter the unfavorable background silica noise arising from the imperfect laser-fiber coupling and light-cladding interaction in the HC-ARF, thus to improve the SNR (ratio of the Gaussian fitted Raman peak intensity of the investigated gas I to the standard deviation of the target spectral range N) for a better LOD as well as the spectral resolution, as shown in Fig. 3(a) ∼ 3(b). Compared with the most commonly used spatial filter system, typically consisting of an aspheric lens, a pinhole (core diameter at μm-level), and a collimating lens, this method dramatically reduces the complexity of the FERS system and the difficulty for optics-operation due to the locations of these iris diaphragms are not precisely required. As a matter of fact, the improvement of SNR is partly at the expense of the target Raman signal intensity. However, the noise intensity is reduced by 15-fold, much larger than the signal intensity does, as shown in Fig. 4(a) ∼ 4(b), and this does more benefit to improving the SNR. Moreover, two iris diaphragms are sufficient for spatial filtering with the best collaboration illustrated in Fig. 2, where ID1 with an aperture around 3.1 mm is mainly used to restrict the redundant input laser beam for improving the laser-fiber coupling, otherwise the beam would over-fill the numerical aperture (NA) of the HC-ARF, and minimizing the interfering Raman signal of the unwanted background silica noise; ID2 with an aperture around 0.8 mm works to filter the majority of the parasitic noise, originating from the light-tube interaction in the HC-ARF, that distributes spatially around the gas Raman signal and could be seen as high frequency signals (away from the optical axis). Additionally, an interesting phenomenon was observed in experiments that the noise intensity would keep dropping within a period of nearly 2 hours since the HC-ARF was exposed to the exciting laser light. The reason may be the matters (e.g., dust, water vapor, etc.) in the fiber core were cleared out at the continuous working of the laser light. Therefore, leaving some time for the “warm-up” is necessary before the experiment.
Fig. 3. Structure and effectiveness of the spatial filtering with iris diaphragms. (a) Schematic sketch for the spatial filtering with two iris diaphragms. The Gaussian-like input Raman signal, accompanied by edge fringes raised from the spatially distributed silica noise and indicated as grey circles, is gradually “cleaned” after the iris diaphragm. (b) The background silica Raman signal obtained with a different number of iris diaphragm from one measurement with an input laser power of 1.3 W and integration time of 20 s.
Fig. 4. Raman measurement of O2, N2 in laboratory air with a different number of the iris diaphragm. (a) Raman spectrum of O2, N2 in laboratory air from one measurement. The bulges a1, a2, a3 indicated with highlight color are efficiently eliminated after two iris diaphragms. (b) Raman signal, noise, and SNR changing with the number of the iris diaphragm. The “Signal” represents the Gaussian fitted Raman peak height of N2 at 2327 cm-1, and the “Noise” is the standard deviation of the silent spectral region from 1800 to 1900 cm-1. Each data point was averaged from 100 measurements and normalized. Except for ID1 and ID2, the third iris diaphragm in (b) was inserted between DBS and M4 and did no improvement for SNR, indicating that two iris diaphragms are sufficient for the spatial filtering. All data in (a) and (b) were acquired with an input laser power of 1.3 W and an integration time of 20 s.
3. Theoretical analysis and experiment
3.1 Signal enhancement of Raman scattering
Generally, the intensity of the recorded Raman signal IRR is proportional to several factors that can be simply described in Eq. (1).
Ideally, introducing a mirror at the rear end facet of the HC-ARF can further enhance the Raman signal intensity by four times [42,45] for the forward propagating laser light, and the backscattered Raman signal would transfer back, which means Le and Ω are doubled. However, considering the transmission attenuation of the fiber and the reflectivity of the mirror, the actual enhancement may be discounted, and the detailed process is described in Fig. 5(a) to 5(e).
Fig. 5. The exciting laser light and Raman scattered light propagating in five conditions. (a) The forward propagating laser light and Raman scattered light in HC-ARF. (b) The forward propagating laser light and backscattered Raman light in HC-ARF with mirror. (c) The forward propagating laser light and reflected forward scattered Raman light in HC-ARF with mirror. (d) The backward propagating laser light and forward scattered Raman light in HC-ARF with mirror. (e) The backward propagating laser light and reflected backward scattered Raman light in HC-ARF with mirror. The “forward” for laser light is defined as the propagating direction from 0 to L, the “forward” for Raman scattered light is defined as the propagating direction following the laser light does, and the “backward” is reversed. ILin represents the input laser intensity at the front fiber end, ILout represents the output laser intensity at the rear end, RL and RR are the reflectivities of the mirror for exciting laser light, and Raman scattered light, respectively, L is the physical length of the HC-ARF, z is the point in HC-ARF where the Raman scattering occurs.
Firstly, considering the transmission functions TL(l) of the laser and TR(l) of the Raman signal, the collected Raman intensity IcR at one fiber end can be expressed as Eq. (2) [46,47].
According to Eq. (2), the intensity of directly backscattered Raman signal IRfb, shown in Fig. 5(b), excited by the forward propagating laser light, can be expressed as Eq. (3) [46–49].
The intensity of the reflected forward scattered Raman signal IRffr by the mirror attached to the rear end of the HC-ARF, shown in Fig. 5(c), excited by the forward propagating laser light, can be expressed as Eq. (4) [46–49].
Similarly, the intensity of directly forward scattered Raman signal IRbf, shown in Fig. 5(d), excited by the backward propagating laser light reflected by the mirror attached to the rear end of the HC-ARF, can be expressed as Eq. (5) [46–49].
Correspondingly, the intensity of the reflected backscattered Raman signal IRbbr by the mirror attached to the rear end of the HC-ARF, shown in Fig. 5(e), excited by the backward propagating laser light reflected by the same mirror, can be expressed as Eq. (6) [46–49].
Finally, the collected Raman signal IRtotal at the front fiber end can be express as Eq. (7).
To investigate the actual Raman signal enhancement introduced by the reflecting mirror MHCF, theoretical calculation and experimental demonstration were simultaneously carried out, and the results are illustrated in Fig. 6(a) and 6(b), respectively. Since the transmission losses of the fiber αL, αR and the reflectivities of the mirror RL, RR for the exciting laser light and Raman scattered light are all over 0.97, which means the optical losses can be almost negligible, the calculation result shows an ideal 4-fold enhancement when using a 0.27m-long HC-ARF. In contrast, the practical experiment disagrees that only presents an enhanced factor of 2.9 for the Raman signal and a gain factor of 2.5 for the SNR, which is most probably caused by the unsatisfied laser-fiber coupling efficiency at the location of the mirror and can be further optimized in the future research. However, the experiment results prove that introducing a reflecting mirror at the rear fiber end is effective for Raman signal enhancement and to achieve a better LOD.
Fig. 6. Comparison of the collected Raman intensity using the HC-ARF with and without a mirror at the rear end in a backscattering arrangement. (a) The calculated relative Raman intensity at different fiber lengths, where αL for 532 nm and αR for 607 nm (Q1 of N2 at 2327 cm-1) were set as 81 dB/km, 101 dB/km respectively, referred to [44], RL and RR were both set as 0.99. The inset shows the relatively Raman intensity at 0.27m-long fiber length. (b) Raman spectra of O2, N2 in the laboratory air, which was acquired from one measurement with a 0.27-m-long HC-ARF, an input laser power of 1.3 W, and an integration time of 60 s.
3.2 Improvement for Raman measurement
Except for the progressive experiment setup, several parameters can also be exploited to improve the collected target signal intensity in practical Raman measurement, as shown in Fig. 7(a) and 7(b), respectively. Theoretically, the Raman signal intensity linearly increases with the exciting laser power and integration time rising, whereas the noise intensity synchronously increases by square root with these two parameters due to the fact that the shot noise obeys the Poisson distribution; thus, the SNR presents the same growth pattern with the noise intensity. In practical experiments, the nonlinear fitted parameter c1 is 0.56, c2 is 0.40 that coincides nicely with the theoretical analysis and indicates that the actual gain would not be pronounced by keeping increasing the exciting laser power and integration time once they reach a value.
Fig. 7. Raman signal, noise, and SNR changing with (a) exciting laser power and (b) integration time. The “Signal” represents the Gaussian fitted Raman peak height of N2 at 2327 cm-1, and the “Noise” is the standard deviation of the silent spectral region from 1800 to 1900 cm-1. Each data point was averaged from 30 measurements and normalized.
4. Results and discussion
4.1 Ambient air
To demonstrate the selectivity, sensitivity, and stability of the novel-innovative fiber-enhanced Raman gas sensing system, Raman spectrum of the ambient air in the laboratory (N2 content of 78.01 vol % and O2 content of 20.95 vol %), as displayed in Fig. 8(a) ∼ 8(c), were firstly investigated with an integration time of 60 s (the HC-ARF used in the experiment had been exposed to the laboratory air for a few days to achieve the gas compositions and concentration balance between the fiber core and outside environment). In Fig. 8(a), two prominent vibrational peaks accompanied by the well-visible ro-vibrational peaks are observed at 1555 cm-1 and 2327 cm-1, respectively, that can be assigned to the Q-branches of ambient O2 and N2 [50,51], with the SNR of 7489 and 30201 as well as the corresponding LODs of 83.9 ppm and 77.5 ppm respectively calculated by LOD = 3C / SNR, in which C is the applied gas concentration. In the low wavenumber spectral region, from 50 to 260 cm-1, multiple peaks intertwined with each other are clearly observed that arise from the rotational transitions of ambient O2 and N2. Another prominent Raman peak observed at 3652 cm-1 can be assigned to H2O (water vapor) in ambient air originating from the symmetric stretching of the O-H band, and the complex ro-vibrational peaks of H2O ranging from 3690 to 4080 cm-1 can also be discriminated after 150-times magnification. The Fermi diad of CO2 (around 400 ppm) are observed at 1286 cm-1 (ν-) and 1388 cm-1 (ν+) respectively, and the LOD is determined to be 23.1 ppm with an SNR of 52 for the ν+ vibration. Additionally, a weak peak is observed at 4629 cm-1. We hypothesize this peak would be the second-order Raman signal excited by the Q-branch of ambient N2.
Fig. 8. Fiber-enhanced Raman spectrum of the ambient air in the laboratory. (a) Spectral overview of the ambient air in the laboratory, including O2, N2, CO2, and H2O. (b) The ro-vibrational Raman transitions of O2. The inset shows the effects due to the spin fine structure of O2 in the region of the Q-branch. (c) The ro-vibrational Raman transitions of N2. All spectra were acquired from one measurement with an input laser power of 1.3 W and an integration time of 60 s. In (b) and (c), A grating with 1800 lines per millimeter (blaze wavelength at 750 nm) was applied for achieving a higher spectral resolution (0.04 nm).
To obtain a high-fineness Raman spectrum of ambient air, the multiple ro-vibrational peaks of O2 and N2 are magnified and separately illustrated in Fig. 8(b) and 8(c). For the ambient O2, S-branch of S(1) to S(35) and O-branch of O(3) to O(31) are well observed with the missing peaks for even J (J represents the rotational quantum number) due to the nuclear spin statistical weight and the weak peak observed at 1508 cm-1 is assigned to the naturally occurring 18O16O in ambient air. Moreover, five well-resolved lines are observed in the region of the Q-branch, as shown in the inset in Fig. 8(b), that originate from the effects due to the spin fine structure of O2. Similarly, for the ambient N2, S-branches of S(0) to S(26) and O-branch of O(2) to O(26) are observed, whereas with the peak intensity ratio of 2: 1 for even and odd J-values. Moreover, the Raman peak of 14N15N is clearly observed at 2291 cm-1 overlapped with the O(5) and O(6) branches of 15N2.
The long-term stability of the FERS system is vital for the accurate quantitative analysis of gases, which was tested by continuously monitoring the Raman spectra of N2 in laboratory air, and the results are shown in Fig. 9. The coefficient of variation (CV, the ratio of the standard deviation to the average) for Gaussian fitted Raman peak height of N2 at 2327 cm-1 varies ∼ 0.27% over a period of 8 hours, whereas the standard deviation of noise shows a significant variation, up to ∼ 0.95%, that is probably caused by the instrument-performance fluctuation (such as the dark current of the CCD), and consequently, the SNR varies ∼ 1.05%. However, all these variations are acceptable in practical application. This benefits from that the HC-ARF is fixed onto an aluminum sheet and with both ends glued to the V-groove holders, which vastly reduce the negative influence on fiber transmission property caused by the mechanical disturbance on HC-ARF.
Fig. 9. Results of the long-term stability test for the FERS system. The “Signal” represents the Gaussian fitted Raman peak height of N2 at 2327 cm-1, and the “Noise” is the standard deviation of the silent spectral region from 1800 to 1900 cm-1. Each data point was averaged from 30 measurements, and the maximum value was normalized to 1. The input laser power was 1.3 W and the integration time was 20 s.
4.2 Hydrogen
The analytical capability of the fiber-enhanced Raman gas sensing system was also evaluated with a standard gas sample of H2 (5000 ppm H2 balanced by Argon with a total pressure of 3.5 bar) and the Raman spectrum, illustrated in Fig. 10(a) and 10(b), were obtained with an integration time of 200 s. In the low wavenumber spectral region, five rotational peaks are well-observed, that S0(0), S0(2), and S0(4) belong to ortho-H2, whereas S0(1) and S0(3) represent para-H2. Correspondingly, in the high wavenumber region, the vibrational and ro-vibrational Raman transitions are predominated, which contribute to more peaks appearing, including the Q1-branch of Q1(0) to Q1(6) with an intensity ratio of 1 : 3 for even and odd J-values determined by the molecular thermal population of rotational ground-state energy levels, S1-branch of S1(0) to S1(3) and O1-branch of O1(2) to O1(3). Since the most intensity Raman peak for H2 is Q1(1), located at 4156 cm-1, and the noise intensity (calculated by the standard deviation of the silent spectral region from 4170 to 4270 cm-1 and averaged from three measurements) is approximately 1.5 times lower than that for S0(1), the peak height of Q1(1) is selected for quantification with an SNR of 2799 and the corresponding LOD of 5.4 ppm.
Fig. 10. Fiber-enhanced Raman spectra of (a) the S0-branch of H2 and (b) the Q1-, S1-, and O1-branches of H2. All spectra were acquired from one measurement with an input laser power of 1.3 W and an integration time of 200 s.
For quantification analysis, a series of standard gas samples of H2 (20 ppm, 100 ppm, 500 ppm, 2500 ppm, and 10000 ppm, respectively, Argon-buffered with a total pressure of 3.5 bar) were investigated with the dynamic range over 2 orders of magnitude. The Raman peak height of Q1(1) at 4156 cm-1 is linearly increasing with the gas concentrations, as shown in Fig. 11, demonstrating the feasibility for accurate gas quantification with the developed FERS system.
Fig. 11. The excellent linearity of fiber-enhanced Raman signal of H2 (peak height of Q1(1) at 4156 cm-1) with rising gas concentration. The background noise intensity is marked as a blue dash line. Each data point was averaged from three measurements with an input laser power of 1.3 W and an integration time of 200 s.
4.3 Sulfur dioxide
The performance of this fiber-enhanced Raman gas sensing system was also evaluated with a standard gas sample of SO2 (5000 ppm SO2 balanced by Argon with a total pressure of 3.5 bar), and the Raman spectra, shown in Fig. 12, was obtained with an integration time of 200 s. Totally, five peaks are observed at 519 cm-1, 859 cm-1, 1150 cm-1, 1286 cm-1, and 1388 cm-1, respectively in the interested spectral region. The former three peaks can be identified as the ν2, (ν3 - ν2), and ν1 vibrations of SO2, whereas the latter two peaks are coincidently the same as the Fermi diad of CO2, which could be the impurity mixed in SO2. Since the most intensity peak for SO2 is v1 band located at 1150 cm-1, the peak height of v1 is selected for quantification with an SNR of 2151 (the noise intensity is calculated by the standard deviation of the silent spectral region from 1160 to 1260 cm-1 and averaged from three measurements) and the corresponding LOD of 7.0 ppm.
Fig. 12. Fiber-enhanced Raman spectrum of SO2. The spectrum was acquired from one measurement with an input laser power of 1.3 W and an integration time of 200 s.
Similarly, for quantification analysis, a series of standard gas samples of SO2 (20 ppm, 100 ppm, 500 ppm, 2500 ppm, and 10000 ppm, respectively, Argon-buffered with a total pressure of 3.5 bar) were investigated with the dynamic range over 2 orders of magnitude. The Raman peak height of v1 at 1150 cm-1 also shows excellent linearly with the gas concentrations increasing, as shown in Fig. 13, which demonstrates the feasibility for accurate gas-quantification with the developed FERS system.
Fig. 13. The excellent linearity of fiber-enhanced Raman signal of SO2 (peak height of v1 at 1150 cm-1) with rising gas concentration. The background noise intensity is marked as a blue dash line. Each data point was averaged from three measurements with an input laser power of 1.3 W and an integration time of 200 s.
In this paper, the 5000 ppm SO2 as well as H2 was filled into FA1 with a differential pressure ∼ 1 bar between the gas cylinder and FA1 until a 3.5-bar gas was delivered (FA1 was at ambient pressure filled with air before experiments). The pressure was roughly controlled via a reducing valve connected to the gas cylinder to permit the gas could be successfully filled into FA1 and simultaneously to avoid a strong airflow (if a large pressure applied) that would disturb the HC-ARF. The response time was around 21 s when the Raman signal reaching 90% of the highest intensity, which means most of the gases has been delivered into the HC-ARF. We also carried out a diffusion experiment, by filling 0.05-bar SO2 (5000 ppm) into a sealed gas chamber where the HC-ARFs were placed and initially filled with air, and results showed that the microchannels could speed up the gas filling process by nearly 13-fold with a response time around 90 s.
5. Conclusion
In this paper, we present an innovative fiber-enhanced Raman gas sensing system using the fabricated 0.27-m-long HC-ARF gas cell, which is attached with a reflecting mirror at the rear end facet that provides a Raman signal enhancement over 2.9 times and a gain factor over 2.5 for SNR than the typical FERS methodology. The excellent performance of the system for multigas analysis is thoroughly demonstrated, thanks to the two iris diaphragms optimized in diameters for high-efficiency spatial filtering that improves the SNR up to 5-fold, by recording Raman spectra of ambient air along with their isotopic molecules in the laboratory, achieving the LODs of 23.1 ppm for CO2, 83.9 ppm for O2, and 77.5 ppm for N2 at 1 bar and an integration time of 60 s, as well as the standard gas samples of H2 and SO2 with the LODs of 5.4 ppm and 7.0 ppm, respectively, at 3.5 bar and an integration time of 200 s. Moreover, the long-term stability of the system is investigated with a maximum CV value of 1.05% for SNR, and six microchannels are drilled along the fiber side that proves the FERS technology with microchannel-modified HC-ARF is feasible and provides potential applications in real-time gas sensing cases that are absent of a pressure differential. Additionally, the excellent linear relationship between Raman signal intensity (peak height) and gas concentrations demonstrates the feasibility of accurate gas quantification. In summary, the introduced FERS system indicates a promising potential for applications in trace gas sensing.
Funding
National Natural Science Foundation of China (U1766217).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. 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,” Sensors and Actuators B: Chemical 232, 188–194 (2016). [CrossRef]
2. L. Dong, F. K. Tittel, C. G. Li, N. P. Sanchez, H. P. Wu, C. T. Zheng, Y. J. Yu, A. Sampaolo, and R. J. Griffin, “Compact TDLAS based sensor design using interband cascade lasers for mid-IR trace gas sensing,” Opt. Express 24(6), A528–A535 (2016). [CrossRef]
3. R. Y. Cui, L. Dong, H. P. Wu, W. G. Ma, L. T. Xiao, S. T. Jia, W. D. Chen, and F. K. Tittel, “Three-dimensional printed miniature fiber-coupled multipass cells with dense spot patterns for ppb-level methane detection using a Near-IR diode laser,” Anal. Chem. 92(19), 13034–13041 (2020). [CrossRef]
4. Y. F. Ma, R. Lewicki, M. Razeghi, and F. K. Tittel, “QEPAS based ppb-level detection of CO and N2O using a high power CW DFB-QCL,” Opt. Express 21(1), 1008–1019 (2013). [CrossRef]
5. H. P. Wu, A. Sampaolo, L. Dong, P. Patimisco, X. L. Liu, H. D. Zheng, X. K. Yin, W. G. Ma, L. Zhang, W. B. Yin, V. Spagnolo, S. T. Jia, and F. K. Tittel, “Quartz enhanced photoacoustic H2S gas sensor based on a fiber-amplifier source and a custom tuning fork with large prong spacing,” Appl. Phys. Lett. 107(11), 111104 (2015). [CrossRef]
6. H. P. Wu, L. Dong, H. D. Zheng, Y. J. Yu, W. G. Ma, L. Zhang, W. B. Yin, L. T. Xiao, S. T. Jia, and F. K. Tittel, “Beat frequency quartz-enhanced photoacoustic spectroscopy for fast and calibration-free continuous trace-gas monitorin,” Nat. Commun. 8(1), 15331 (2017). [CrossRef]
7. S. Z. Li, L. Dong, H. P. Wu, A. Sampaolo, P. Patimisco, V. Spagnolo, and F. K. Tittel, “Ppb-Level Quartz-Enhanced Photoacoustic Detection of Carbon Monoxide Exploiting a Surface Grooved Tuning Fork,” Anal. Chem. 91(9), 5834–5840 (2019). [CrossRef]
8. K. Chen, M. Gao, S. Liu, B. Zhang, H. Deng, Y. H. Zheng, Y. W. Chen, C. Luo, L. Tao, and M. Q. Lou, “Fiber-optic photoacoustic sensor for remote monitoring of gas micro-leakage,” Opt. Express 27(4), 4648–4659 (2019). [CrossRef]
9. Z. F. Gong, T. L. Gao, Y. W. Chen, B. Zhang, W. Peng, Q. X. Yu, F. X. Ma, L. Mei, and K. Chen, “Sub-ppb level detection of nitrogen dioxide based on an optimized H-type longitudinal acoustic resonator and a lock-in white-light interferometry demodulation algorithm,” Journal of Quantitative Spectroscopy and Radiative Transfer 253, 107136 (2020). [CrossRef]
10. Z. F. Gong, G. J. Wu, X. Jiang, H. Li, T. L. Gao, M. Guo, F. X. Ma, K. Chen, L. Mei, W. Peng, and Q. X. Yu, “All-optical high-sensitivity resonant photoacoustic sensor for remote CH4 gas detection,” Opt. Express 29(9), 13600–13609 (2021). [CrossRef]
11. T. H. Yang, W. G. Chen, Z. X. Zhang, J. L. Lei, F. Wan, and R. M. Song, “A multiple reflections enhanced fiber-optic photoacoustic sensor for gas micro-leakage,” Opt. Express 29(2), 2142–2152 (2021). [CrossRef]
12. Y. F. Ma, Y. He, Y. Tong, X. Yu, and F. K. Tittel, “Quartz-tuning-fork enhanced photothermal spectroscopy for ultra-high sensitive trace gas detection,” Opt. Express 26(24), 32103–32110 (2018). [CrossRef]
13. Y. He, Y. F. Ma, Y. Tong, X. Yu, and F. K. Tittel, “Ultra-high sensitive light-induced thermoelastic spectroscopy sensor with a high Q-factor quartz tuning fork and a multipass cell,” Opt. Lett. 44(8), 1904–1907 (2019). [CrossRef]
14. Y. F. Ma, Y. He, P. Pastimisco, A. Sampaolo, S. D. Qiao, X. Yu, F. K. Tittel, and V. Spagnolo, “Ultra-high sensitive trace gas detection based on light-induced thermoelastic spectroscopy and a custom quartz tuning fork,” Appl. Phys. Lett. 116(1), 011103 (2020). [CrossRef]
15. F. M. Schmidt, A. Foltynowicz, W. G. Ma, and O. Axner, “Fiber-laser-based noise-immune cavity-enhanced optical heterodyne molecular spectrometry for Doppler-broadened detection of C(2)H(2) in the parts per trillion range,” J. Opt. Soc. Am. B 24(6), 1392–1405 (2007). [CrossRef]
16. G. Zhao, T. Hausmaninger, F. M. Schmidt, W. G. Ma, and O. Axner, “High-resolution trace gas detection by sub-Doppler noise-immune cavity-enhanced optical heterodyne molecular spectrometry: application to detection of acetylene in human breath,” Opt. Express 27(13), 17940–17953 (2019). [CrossRef]
17. Y. T. Zhou, J. X. Liu, S. J. Guo, G. Zhao, W. G. Ma, Z. S. Cao, L. Dong, L. Zhang, W. B. Yin, Y. Q. Wu, L. X. Xiao, O. Axner, and S. T. Jia, “Laser frequency stabilization based on a universal sub-Doppler NICE-OHMS instrumentation for the potential application in atmospheric lidar,” Atmos. Meas. Tech. 12(3), 1807–1814 (2019). [CrossRef]
18. M. Hippler, “Cavity-enhanced Raman spectroscopy of natural gas with optical feedback cw-diode lasers,” Anal. Chem. 87(15), 7803–7809 (2015). [CrossRef]
19. T. W. Smith and M. Hippler, “Cavity-enhanced Raman spectroscopy in the biosciences: In situ, multicomponent and isotope selective gas measurements to study hydrogen production and consumption,” Anal. Chem. 89(3), 2147–2154 (2017). [CrossRef]
20. A. Siburg, T. Jochum, S. E. Trumbore, J. Popp, and T. Frosch, “Onsite cavity enhanced Rama spectrometry for the investigation of gas exchange processes in the earth’s critical zone,” Analyst 142(18), 3360–3369 (2017). [CrossRef]
21. A. Sieburg, S. Schneider, D. Yan, J. Popp, and T. Frosch, “Monitoring of gas composition in a laboratory biogas plant using cavity enhanced Raman spectroscopy,” Analyst 143(6), 1358–1366 (2018). [CrossRef]
22. P. Y. Wang, W. G. Chen, J. X Wang, J. Tang, Y. L. Shi, and F. Wan, “Multigas analysis by cavity-enhanced Raman spectroscopy for power transformer diagnosis,” Anal. Chem. 92(8), 5969–5977 (2020). [CrossRef]
23. D. V. Petrov, “Multipass optical system for a Raman gas spectrometer,” Appl. Opt. 55(33), 9521–9525 (2016). [CrossRef]
24. J. G. Velez and A. Muller, “Trace gas sensing using diode-pumped collinearly detected spontaneous Raman scattering enhanced by a multipass cell,” Opt. Lett. 45(1), 133–136 (2020). [CrossRef]
25. J. G. Velez and A. Muller, “Spontaneous Raman scattering at trace gas concentrations with a pressurized external multipass cavity,” Meas. Sci. Technol. 32(4), 045501 (2021). [CrossRef]
26. W. F. Pearman, J. C. Carter, S. M. Angel, and J. W. J. Chen, “Multipass capillary cell for enhanced Raman measurements of gases,” Appl. Spectrosc. 92(8), 5969–5977 (2008). [CrossRef]
27. W. F. Pearman, J. C. Carter, S. M. Angel, and J. W. J. Chen, “Quantitative measurements of CO2 and CH4 using a multipass Raman capillary cell,” Appl. Opt. 47(25), 4627–4632 (2008). [CrossRef]
28. M. P. Buric, K. P. Chen, J. Falk, and S. D. Woodruff, “Multimode metal-lined capillaries for Raman collection and sensing,” J. Opt. Soc. Am. B 27(12), 2612–2619 (2010). [CrossRef]
29. M. P. Buric, K. P. Chen, J. Falk, and S. D. Woodruff, “Enhanced spontaneous Raman scattering and gas composition analysis using a photonic crystal fiber,” Appl. Opt. 47(23), 4255–4261 (2008). [CrossRef]
30. M. P. Buric, K. P. Chen, J. Falk, and S. D. Woodruff, “Improved sensitivity gas detection by spontaneous Raman,” Appl. Opt. 48(22), 4424–4429 (2009). [CrossRef]
31. S. Hanf, R. Keiner, D. Yan, J. Popp, and T. Frosch, “Fiber-enhanced Raman multigas spectroscopy: a versatile tool for environmental gas sensing and breath analysis,” Anal. Chem. 86(11), 5278–5285 (2014). [CrossRef]
32. S. Hanf, T. Bogozi, R. Keiner, T. Frosch, and J. Popp, “Fast and highly sensitive fiber-enhanced Raman spectroscopic monitoring of molecular H2 and CH4 for point-of-care diagnosis of malabsorption disorders in exhaled human breath,” Anal. Chem. 87(2), 982–988 (2015). [CrossRef]
33. T. Jochum, L. Rahal, R. J. Suckert, J. Popp, and T. Frosch, “All-in-one: a versatile gas sensor based on fiber enhanced Raman spectroscopy for monitoring postharvest fruit conservation and ripening†,” Analyst 141(6), 2023–2029 (2016). [CrossRef]
34. A. Knebl, R. Domes, D. Yan, J Popp, S. Trumbore, and T. Frosch, “Fiber-enhanced Raman gas spectroscopy for O-18-C-13-labeling experiments,” Anal. Chem. 91(12), 7562–7569 (2019). [CrossRef]
35. A. Knebl, R. Domes, S. Wolf, C. Domes, J Popp, and T. Frosch, “Fiber-enhanced Raman gas spectroscopy for the study of microbial methanogenesis,” Anal. Chem. 92(18), 12564–12571 (2020). [CrossRef]
36. A. Knebl, C. Domes, R. Domes, S. Wolf, J. Popp, and T. Frosch, “Hydrogen and C2−C6 alkane sensing in complex fuel gas mixtures with fiber-enhanced Raman spectroscopy,” Anal. Chem. 93(30), 10546–10552 (2021). [CrossRef]
37. Y. L. Hoo, W. Jin, C. Z. Shi, H. L. Ho, D. N. Wang, and S. C. Ruan, “Design and modeling of a photonic crystal fiber gas sensor,” Appl. Opt. 42(18), 3509–3515 (2003). [CrossRef]
38. M. W. Rosalind, B. Banafsheh, J. G. Kevin, and O. Alfonso, “Sub-minute response time of a hollow-core photonic bandgap fiber gas sensor,” IEEE Photon. Technol. Lett. 27(11), 1590–1596 (2009). [CrossRef]
39. Y. L. Hoo, S. J. Liu, H. L. Ho, and W. Jin, “Fast response microstructured optical fiber methane sensor with multiple side-openings,” IEEE Photonic. Tech. L. 22(5), 296–298 (2010). [CrossRef]
40. F. Yang, W. Jin, Y. C. Lin, C. Wang, H. Lut, and Y. Z. Tan, “Hollow-core microstructured optical fiber gas sensors,” J. Lightwave Technol. 35(16), 3413–3424 (2017). [CrossRef]
41. P. C. Zhao, Y. Zhao, H. H. Bao, H. L. Ho, W. Jin, S. C. Fan, S. F. Gao, Y. 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]
42. Q. Chu, Z. Q. Jin, X. T. Yu, C. X. Li, W. H. Zhang, W. B. Ji, B. Lin, P. P. Shum, X. P. Zhang, and G. H. Wang, “Volumetric enhancement of Raman scattering for fast detection based on a silver-lined hollow-core fiber,” Opt. Express 27(7), 10370–10382 (2019). [CrossRef]
43. X. T. Yu, C. X. Li, D. J. J. Hu, K. Milenko, G. H. Wang, P. Shum, F. Xu, Y. Q. Lu, and X. P. Zhang, “All-fiber online Raman sensor with enhancement via a Fabry–Perot cavity,” Opt. Lett. 45(20), 5760–5763 (2020). [CrossRef]
44. S. F. Gao, Y. Y. Wang, X. L. Liu, C. Hong, S. Gu, and P. Wang, “Nodeless hollow-core fiber for the visible spectral range,” Opt. Lett. 42(1), 61–64 (2017). [CrossRef]
45. Z. Q. Jin, Q. Chu, W. H. Xu, H. Cai, W. B. Ji, G. H. Wang, B. Lin, and X. P. Zhang, “All-Fiber Raman Biosensor by Combining Reflection and Transmission Mode,” IEEE Photonics Technol. Lett. 30(4), 387–390 (2018). [CrossRef]
46. A. Robert, D. M. Michelle, P. V. D. Richard, and K. Ilia, “Intensity considerations in liquid core optical fiber Raman spectroscopy,” Appl. Spectrosc. 55(4), 373–381 (2001). [CrossRef]
47. T. Frosch, D. Yan, and J. Popp, “Ultrasensitive fiber enhanced UV resonance Raman sensing of drugs,” Anal. Chem. 85(13), 6264–6271 (2013). [CrossRef]
48. D. Yan, J. Popp, and T. Frosch, “Analysis of fiber-enhanced Raman gas sensing based on raman chemical imaging,” Anal. Chem. 89(22), 12269–12275 (2017). [CrossRef]
49. A. Knebl, D. Yan, J. Popp, and T. Frosch, “Fiber enhanced Raman gas spectroscopy,” Anal. Chem. 103, 230–238 (2018). [CrossRef]
50. T. Jochum, A. Fastnacht, S. E. Trumbore, J. Popp, and T. Frosch, “Direct Raman spectroscopic measurements of biological nitrogen fixation under natural conditions: an analytical approach for studying nitrogenase activity,” Anal. Chem. 89(2), 1117–1122 (2017). [CrossRef]
51. A. Sieburg, A. Knebl, J. M. Jacob, and T. Frosch, “Characterization of fuel gases with fiber-enhanced Raman spectroscopy,” Anal. Bioanal. Chem. 411(28), 7399–7408 (2019). [CrossRef]
