Open Access
Add to BibSonomy
Abstract
A cavity-enhanced Raman spectroscopy (CERS) gas-sensing method is introduced. Using optical feedback frequency-locking, laser radiation provided by a diode laser is coupled into a three-mirror V-shaped optical cavity. An intracavity laser power of 92 W is realized, yielding a power gain factor of 2200. Raman spectrums of air, carbon dioxide, and acetylene are recorded as a demonstration. Multicomponent gas mixtures including isotopic gases can be simultaneously sensed by CERS. With 200 s exposure time, detection limits of 5.35 Pa for N2, 5.07 Pa for O2, 1.74 Pa for CO2, and 0.58 Pa for C2H2 are achieved. CERS is a powerful gas-sensing method with high selectivity and sensitivity, which also has the potential for quantitative analysis of gases with high accuracy.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Spectroscopic technology is an indispensable method for gas sensing due to the advantages of non-invasiveness, high selectivity, no sample consumption, and allowing in-situ real-time monitoring. Commonly used spectroscopic approaches for gas sensing are infrared absorption-based techniques, such as direct infrared absorption spectroscopy [1–3] or photoacoustic spectroscopy [4–9]. They provide excellent sensitivity because of the high absorption cross-sections. However, diatomic homonuclear gases, which are important characteristic gases in many fields, are hard to be detected by infrared absorption-based techniques. Besides, different laser wavelengths are required for different gases, which means the sensing of gas mixtures may need several laser sources and detectors.
Raman spectroscopy [10,11] is an alternative powerful method for gas sensing. Almost all gases can be detected except for monoatomic gases; multicomponent gases can be measured simultaneously using a single-wavelength laser and the same detector. Unfortunately, trace gases are difficult to be sensed by Raman spectroscopy due to the insufficient scattering intensity resulting from the extremely low Raman scattering cross-sections of gases [12,13]. The intensity of Raman scattering scales linearly with laser power, so an optical resonant cavity can help overcome the weak signal of Raman scattering. For a high-finesse optical cavity, if the resonance condition is satisfied (cavity length is an integral multiple of half laser wavelength), the laser power can be built up by several orders of magnitude because of multiple-beam constructive interference (resonance). Therefore, the cavity-enhanced technologies are very beneficial for Raman signal enhancement. However, the resonance condition is challenging to be maintained because the cavity length cannot remain constant for a long time. Mechanical waves or temperature shifts can easily affect the cavity length. Accordingly, an appropriate frequency-locking method that maintains the resonance is necessary for cavity-enhanced technologies.
One approach to achieve frequency-locking is the active method, such as the Pound-Drever-Hall (PDH) technique [14–18]. The difference between laser frequency and cavity resonance frequency can be calculated by the PDH system in real-time, then the cavity length (cavity resonance frequency) can be adjusted actively by a piezo actuator (PZT) to match the laser frequency. Friss et al. [19] achieved a PDH frequency-locking CERS system. With 3.7 mW input power, the intracavity laser power reaches 22 W, yielding a power gain factor of 5900.
The laser provided by a semiconductor laser can be passively locked into an optical cavity based on the optical feedback frequency-locking (OFFL) [20–36]. The optical cavity also serves as a frequency filter, the laser radiation whose frequency is exactly the same as the cavity resonance frequency can be established and exist in the cavity. If the laser beam directly reflected by the cavity mirror is eliminated by an appropriate way and only the light output from the cavity feeds back to laser source, then according to optical injection locking [37,38], the frequency of the feedback light, which is same as the cavity resonance frequency, will be copied to the laser source, thus resulting a huge laser power buildup in the cavity. Hippler et al. [20] developed an OFFL CERS system in which the direct reflection beam is eliminated by optical isolators. An intracavity laser power of 2.5 W with a power gain factor of 833 is achieved. Popp et al. [21–28] described an OFFL CERS system in which the elimination of the direct reflection beam is realized by spatial filtering [21]. An intracavity laser power of 100 W with a power gain factor of 2000 is achieved. This system has been successfully used in many fields, such as environmental research [21–24], microbiology research [25,26], and botanical research [27,28].
The intensity of Raman scattering can be further enhanced in a microcavity due to the improvement of spontaneous emission (Purcell effect) [39–42]. Petrak et al. [43] developed a CERS system based on a microcavity with PDH frequency-locking technique. To realize the small cavity volume, the microcavity is consisted of a micromirror on a planar substrate facing a nominally identical mirror at the tip of a single-mode fiber. The size of this microcavity is around 100 µm3. However, the cavity length of the microcavity is too short (micron scale), which results in an insufficient laser-gas interaction length. Therefore, the limit of detection (LOD) is not as good as expected.
In this work, we introduce a cavity-enhanced Raman gas-sensing apparatus based on optical feedback frequency locking, in which direct reflection beam is eliminated by the V-shaped cavity configuration. A 642 nm diode laser is coupled into a three-mirror V-shaped optical cavity. With 42 mW input power, an intracavity laser power of 92 W is achieved, yielding a power gain factor of 2200. To get stronger scattering intensity, Raman-scattered light is collected in a forward scattering collection geometry. The performance of the apparatus is demonstrated by recording the Raman spectrums of air, carbon dioxide, and acetylene. The results show that CERS is a powerful gas-sensing method with high sensitivity and selectivity, which is also feasible for quantitative analysis of gases with high accuracy.
2. Experimental setup
The experimental setup is shown in Fig. 1. A diode laser (DL, Beijing Laserwave LWRL642) which emits linear polarization radiation at 642 nm is used as the light source. A bandpass clean-up filter (BF, 636-644 nm, Semrock LD01-640/8) blocks the unwanted broad emission around 720 nm output from the DL. The V-shaped cavity is formed by three high-reflectivity cavity mirrors (CM1, CM2, CM3, Advanced Thin Films, customized). CM1 and CM2 have a 1 m radius of curvature, CM3 is a plane mirror. At 642 nm the reflectivity of CM1, CM2, and CM3 at 0° incidence angle R = 99.992% has been determined by cavity-ring down measurement. The length of both cavity-arms is L = 500 mm, and the angle between the two arms is 4°.
Fig. 1. Schematic diagram of the V-shaped cavity-enhanced Raman spectroscopy setup. More details are presented in the main text.
An achromatic half-wave plate (HP1, Thorlabs AHWP05M-600) located behind the optical isolator (OI, 30dB at 642 nm, Thorlabs IO-3D-633-PBS) is used to adjust the polarization of laser inside the cavity. Mode-matching lens (ML, focus length f = 500 mm) couples the laser beam into the cavity. Another achromatic half-wave plate (HP2) is used to adjusted the polarization of the scattered light output from CM2, after that the light which has the same polarization with intracavity laser passes through the polarizing beam splitter (PBS, Thorlabs CCM1-PBS252/M). The scattered light output from CM1 which has the same polarization with intracavity laser is reflected by the PBS simultaneously. Therefore, the scattered light output from both arms can propagate along the same axis after PBS. Three long-pass edge filters (EF, the cut-off wavelength of 671 nm, which corresponds to 673 cm−1 with 642 nm Raman excitation, Semrock LP02-671RU) sufficiently block the laser radiation and Rayleigh scattering to suppress the spectral noise. To monitor the cavity output signal, the first EF is slightly angled so the light can be reflected into a photodetector (PD, Thorlabs PDA100A-EC). Raman-scattered light passes the EFs and is focused by an achromatic lens into the monochromator (Princeton Instruments HRS-300S) and recorded by a CCD camera (Princeton Instruments PIX-400B). The grating used in this experiment is blazed at 750 nm with 1200 lines per mm. The gas cell has three optical windows coated with antireflective film, and they are slightly angled to avoid direct laser reflection. A pressure sensor is mounted on the gas cell to monitor the gas pressure inside the cell. Gases in the cell can be pumped out by a vacuum pump (oil-free vacuum pump is used to avoid the pollution of cavity mirrors caused by oil mist). A large-size glass window is also mounted on the cell cover to observe the intracavity laser beams.
The optical cavity also acts as a frequency filter, only the laser radiation whose frequency is exactly the same as the cavity resonance frequency can be established in the cavity, then it will output and feedback to the DL. The light directly reflected by CM3 does not return to the DL due to the V-shaped configuration, and it is blocked by a beam trap to ensure safety. According to the injection locking of the diode laser, frequency-locking and consequent power buildup always occur in the V-shaped cavity after carefully adjusting the cavity mirrors. However, in that case, the intracavity laser radiation occurs with reduced power and broad bandwidth, which is not conducive to Raman gas sensing with high sensitivity and selectivity. Intracavity laser power can be improved by adjusting the feedback phase and intracavity laser bandwidth can be narrowed by reducing the feedback rate (the ratio of feedback power to laser output power).
The feedback phase, which depends on the laser-cavity distance between the DL and the CM3, has a great impact on the intracavity laser power [30]. For a V-shaped cavity, the feedback phase is optimal when laser-cavity distance Lc is an exactly odd-multiple of the cavity arm length L [36]. In our experiment, the Lc is roughly adjusted to Lc = 3L first by a displacement stage (DS, OptoSigma TSD-60121S) mounted under the DL. For the precise adjustment of laser-cavity distance (feedback phase), if the laser frequency is periodically linearly modulated in one direction, the profile of the cavity output signal can be used to evaluate whether the feedback phase is optimum on the nanoscale [31]. Then, the laser-cavity distance can be controlled by a PZT (Physik Instrumente P-725K078) mounted under the M2. To achieve this, the injection current (laser frequency) of the DL is modulated by a function generator (Tektronix AFG3022C) using a sawtooth wave at 2 kHz. The cavity output signal in one laser frequency modulation period is recorded by a PD and transfers to a data acquisition card (DAQ, National Instruments USB-6259) during the PZT movement. As shown in Fig. 2(a), the symmetry of the cavity output signal in one laser frequency modulation period can be used to evaluate whether the feedback phase is correct because when the feedback phase is optimum, the shape of the curve is symmetrical [31]. Specifically, a feedback loop, designed as a LabVIEW program based on the method described in [32], reduces noise, differentiates and then integrates the cavity output signal to obtain an error signal which is proportional to the phase error. The PZT is driven according to this error signal to correct the laser-cavity distance (feedback phase).
Fig. 2. Cavity output signals and DL output signals in one laser frequency modulation period. (a) Cavity output signals. The shape of the signal is asymmetric when the feedback phase is incorrect. A PZT is controlled by an error signal to adjust the laser-cavity distance to achieve an optimal feedback phase, which is presented as a symmetrical cavity output signal. (b) DL output signals. When the cavity is absent, the waveform of the DL output signal is the same as the modulation signal. When the cavity is present, the gain of DL is affected by the optical feedback, and a large laser power increase is observed when the feedback phase is optimum. Signals with incorrect feedback phase are acquired by keeping PZT moving in one direction.
As a result, in each modulation period, the laser frequency changes by the current modulation until it is locked to a longitudinal mode of the cavity with the optimal feedback phase. As the modulation continues, this resonance is then lost after some time and appears again in the next period. By choosing an appropriate offset and amplitude of modulation signal, the resonances occur in the middle of each period with a duty cycle of about 50%, as shown in Fig. 2(a). A trigger signal also outputs from the function generator and transfers to itself and the DAQ, so that the cavity output signal acquisition and the laser frequency modulation can be synchronized.
To observe the output signals of the DL with different feedback phases, a laser power sampling mirror is temporarily placed between M2 and OI, about 10% of laser power can be reflected into another photodetector to monitor the DL output signal. The output power of DL is enhanced or suppressed in the presence of cavity feedback as shown in Fig. 2(b), because the gain of the laser diode is affected by the optical feedback, increasing or decreasing depends on the feedback phase [33–35]. A large laser power increase is observed when the feedback phase is optimum. In other experiments, the sampling mirror is removed to achieve a stronger intracavity laser power.
The feedback rate is another important parameter. For the high feedback rate, the laser frequency may jump from one resonance to another [36], it indicates a wide linewidth of intracavity laser, which is disadvantageous to high selectivity sensing of multicomponent gases by Raman spectroscopy. Therefore, it is necessary to reduce the feedback rate so that the frequency jumping between several longitudinal modes can be avoided [30]. To achieve this, an OI which is located between the cavity and the DL reduces the feedback rate to about 5×10−5. As a result of laser frequency modulation and feedback rate reduction, always one same longitudinal mode is excited in the cavity, thus the intracavity laser bandwidth can be narrowed and the spectral resolution can be improved.
The laser power after ML is measured as Iin = 42.0 mW, and the average intracavity laser power in both arms Ic = 80 W is calculated from
where Iout is the output laser power measured as 3.19 mW after both CM1 and CM2. The factor of 1 + R contributes from the laser power traveling in both directions along the cavity axes [44].If the optical cavity is placed in the closed gas cell, the output laser power will raise to 3.55 mW after standing 48 hours or raise to 3.86 mW after air pumping and then fall to 3.70 mW (Ic = 92 W with power gain factor of 2200) if gas cell is filled with sample gases, as shown in Fig. 3. This phenomenon is attributed to the fact that intracavity laser power is affected by the density of the atmospheric particulate matters. As shown in Fig. 4, the larger particulate matters in the air such as dust will slowly deposit on the bottom of the gas cell if the cell is airtight. Therefore, the intracavity loss, which is positively correlated with air dust density between cavity mirrors, decreases with the standing time. Similarly, the particulate matters will be pumped out during air pumping or enter the cell again during sample gas filling, which results in the change of intracavity laser power.
Fig. 3. The changes of cavity output laser power with standing time or air pumping. The cavity output power raises to 3.55 mW from 3.19 mW after the enhanced-cavity has been stood for 48 hours in the closed gas cell, and returns to 3.19 mW when the gas cell cover is removed. After air pumping, the cavity output power increases to 3.86 mW, and falls to 3.70 mW when the gas cell is filled with sample gases. Finally, it backs to 3.19 mW when pumping out the sample gases and opening the valve of gas cell.
Fig. 4. Photographs of intracavity laser beams, which are photographed through the large-size observation window mounted on the gas cell cover. (a) Without standing or air pumping. Many atmospheric particulate matters such as dust can be seen in the laser beams. The beams look brightest as a result of the strong Mie scattering or Tindall effect but have the weakest intensity. Iout is measured as 3.19 mW and Ic is calculated as 80 W. (b) After standing 48 hours. The beams look cleaner and have a higher intensity. Iout is measured as 3.55 mW and Ic is calculated as 88 W. (c) After air pumping. The laser beams are hardly visible due to almost all particulate matters in the air are pumped out but has the strongest intensity. Iout is measured as 3.86 mW and Ic is calculated as 97 W. (d) After air pumping and sample gases filling. The laser beams can be seen again because a small amount of particulate matters is contained in the sample gases. Iout is measured as 3.70 mW and Ic is calculated as 92 W.
The polarization of the laser inside the cavity and the angle between two arms also affect the intracavity laser power. According to the Fresnel equations, the reflectivity of CM3 to p-polarized wave is lower than that to s-polarized wave. Therefore, to acquire a stronger intracavity laser power, an achromatic half-wave plate (HP1) is used to adjust the polarization of laser inside the cavity to s-polarized, which means the angle between laser polarization direction and the normal of the incident plane of CM3 α = 0°. In our experiment, the output laser power falls to 2.30 mW from 3.19 mW when the intracavity is adjusted to p-polarized, as shown in Fig. 5(a). The intracavity laser power decreases with the expansion of arms’ angle as shown in Fig. 5(b), the designed incident angle of CM3 is 0°, so CM3 has a lower reflectivity at larger incident angles according to the principle of optical coating of mirrors.
Fig. 5. The cavity output laser power changed with (a) intracavity laser polarization and (b) arms’ angle. α is the angle between laser polarization direction and the normal of the incident plane of CM3. To avoid the intracavity laser power changing with standing time, the optical cavity is not placed in the gas cell in these experiments. Besides, the gas cell has not enough room for a big arms’ angle.
To obtain stronger scattering intensity, Raman signals output from both arms are collected simultaneously based on the polarization of scattered light. For the low-depolarized Raman-scattered light, which means the polarization of scattered light is almost the same as that of the intracavity laser, the polarization of scattered light output from CM2 can be controlled to perpendicular with that output from CM1 by a half-wave plate (HP2), then two beams output from CM1 and CM2 can propagate along the same axis after PBS. Therefore, ideally the scattering intensity can be almost doubled by double-arm collection compared with the single-arm collection. However, the grating diffraction efficiency is usually different for s-polarized wave and p-polarized wave, so the scattering intensity recorded by CCD is generally less than double.
On the other hand, for the highly-depolarized Raman-scattered light, which means the scattered light is almost unpolarized, the scattering intensity will be enhanced slightly by double-arm collection because the “wrong-polarized scattering light” from each arm is wasted by the PBS, as shown in Fig. 6. The Raman scattered-light composes of s-polarized wave and p-polarized wave, and assuming the intensity is Is and Ip respectively. For single-arm collection, the scattering intensity recorded by CCD I1 can be ideally considered as:
where ηs and ηp is the grating diffraction efficiency for s-polarized wave and p-polarized wave respectively, EQ is the quantum efficiency of CCD. For double-arm collection, the scattering intensity recorded by CCD can be ideally considered as:Fig. 6. The comparison of single-arm collection and double-arm collection. (a) Single-arm collection. CCD records both of s-polarized and p-polarized Raman-scattered light. (b) Double-arm collection. The p-polarized Raman-scattered light output from both arms is wasted by PBS, and only s-polarized Raman-scattered light can be recorded. Therefore, the double-arm collection is a useful improvement method for low-depolarized Raman-scattered light, but it is not very effective for highly-depolarized Raman-scattered light.
For example, assuming ηs = 0.8 and ηp = 0.6, if the depolarization ratio of Raman-scattered light is zero (Ip = 0), then I2 / I1 = 1.75 could be calculated, which means the recorded scattering intensity can be improved effectively using double-arm collection for the low-depolarized Raman-scattered light; on the other hand, if the depolarization ratio of Raman-scattered light is 0.75 (Ip = 0.75Is), I2 / I1 = 1.12 could be calculated, which means the recorded scattering intensity is only improved a little for the highly-depolarized Raman-scattered light.
3. Results and discussion
3.1. Air
To demonstrate the performance of the CERS gas-sensing apparatus, Raman spectrums of ambient laboratory air, carbon dioxide, and acetylene were recorded. The Raman spectrums of 100 kPa ambient laboratory air with an exposure time of 20 s and 100 µm slit width are shown in Figs. 7(a)–7(c) (the enhanced-cavity has been stood for 48 hours in the airtight gas cell to achieve a stronger intracavity laser power). The spectrum in Fig. 7(a) is mainly composed of the Q-branch of O2 observed at 1554 cm−1, the Q-branch of N2 observed at 2327 cm−1, and water vapor from air humidity (a broad peak observed between 3500 cm−1 to 4500 cm−1 with strong scattering intensity, which is about 11 times than that of N2). CO2 in the air, O- and S-branches of O2 and N2 can also be clearly presented after zooming in the spectrum. For the N2, S-branches of S (0) to S (20) and O-branches of O (3) to O (22) can be observed, as shown in Fig. 7(b). An overlapped peak (marked by an asterisk) is also observed between the O (5) transition and O (6) transition. For the O2, S-branches of S (1) to S (27) and O-branches of O (5) to O (27) can be observed. For the CO2, the Fermi resonance pair is observed at 1285 cm−1 and 1388 cm−1, as shown in Fig. 7(c). Because of the nuclear spin statistical weight for even and odd J, the vibration-rotation Raman transitions from even J of O2 are missing, and the vibration-rotation Raman transitions of N2 showing an alternation of strength and weakness with 2 : 1 for even and odd J.
Fig. 7. The Raman spectrums of ambient laboratory air. For (a), (b), and (c), the slit width is set as 100 µm. In (d), the slit width is set as 50 µm. (a) An overview spectrum of laboratory air, including O2, N2, CO2, and water vapor from air humidity. (b) The vibration-rotation Raman transitions of N2. The O (5) transition and O (6) transition are overlapped, marked by an asterisk. (c) The vibration-rotation Raman transitions of O2. The O (3) transition of O2 and O (2) transition of N2 are overlapped by their respective Q-branch. (d) The vibration-rotation Raman transitions of N2 with 50 µm slit width.
To figure out the reason why O (5) transition and O (6) transition of N2 are overlapped, the Raman spectrum with 50 µm slit width is recorded as shown in Fig. 7(d). The 14N15N whose line lies between O (5) transition and O (6) transition of 14N2 is also observed [45]. Spectral resolution can be improved by reducing slit width, but the signal intensity will be reduced. The scattering intensity with 50 µm slit width is dropped by nearly two-thirds to that with 100 µm slit width.
To obtain the LOD of gases measured by this apparatus, the signal-to-noise ratio (SNR) needs to be acquired. In this work, signal intensity (I) is considered as the Gaussian fitted Raman peak height [46]. The noise intensity (N) is considered as the average standard deviation of the peak range calculated from 20 blank spectrums, which are acquired with the same exposure time but no detected gases. According to the gas partial pressure (P) and SNR (I / N), LOD is calculated as LOD = 3P / SNR. With 20 s exposure time and 100 µm slit width, for N2 in the air (78 kPa), a SNR of 12793 for its Q-branch and a corresponding LOD of 16.62 Pa are achieved; for O2 in the air (21 kPa), a SNR of 3716 for its Q-branch and a corresponding LOD of 15.43 Pa are achieved; for CO2 (1388 cm−1) in the air, a SNR of 21 and a corresponding LOD of 5.43 Pa are achieved, based on the global average annual concentration of CO2 in the atmosphere is 407.4 ppm in 2018. It should be noted that about 4% of air is retained in the cell after pumping, the Raman peaks of N2 and O2 can still be measured, so the noise intensity of the N2 and O2 is estimated from 1850 cm−1 to 1950 cm−1 in blank spectrums.
To demonstrate the improvement of double-arm collection compared with the single-arm collection, the scattering intensity of gases acquired from single and double arms are summarized in Table 1. The Raman spectrums with single-arm (arm 2) collection are obtained by blocking the beam output from CM1 and removing the PBS and HP2. The scattering intensity of Q-branches of N2 and O2 is obviously improved by double-arm collection, the intensity of O- and S-branches are slightly increased.
Table 1. Comparison of double-arm collection and single-arm collection.
3.2. Carbon dioxide
For other detected gases, the previous gases inside the cell are pumped out, but there will be about 4% (4 kPa) of them still remained in the cell after pumping. With an exposure time of 200 s and a slit width of 100 µm, the Raman spectrum of 100 kPa CO2 standard gas (the concentration of CO2 is 5000 ppm and carried by Ar, the total pressure in the cell is 104 kPa) is shown in Fig. 8. The spectrum is mainly composed of 12CO2 (including the ν1/2ν2 Fermi resonance pair observed at 1285 cm−1 and 1388 cm−1, the hot bands observed at 1265 cm−1 and 1409 cm−1), 13CO2 observed at 1269 cm−1 [47], and the remained O2.
Fig. 8. The Raman spectrum of CO2. The spectrum is mainly composed of CO2 (including 12CO2 and 13CO2) and remained O2.
With 200 s exposure time and 100 µm slit width, a SNR of 866 and a corresponding LOD of 1.74 Pa for CO2 (1388 cm−1) are achieved, based on the partial pressure of CO2 in the cell is 501.63 Pa (500 Pa from the standard gas and 1.63 Pa from the remained air); for Q-branch of remained O2, a SNR of 497 and a corresponding LOD of 5.07 Pa are achieved. It can be seen from the LOD of O2 with different exposure times, the LOD does not linearly improve with the exposure time. Theoretically, the intensity of Raman scattering increases linearly with exposure time, but the noise intensity (standard deviation of noise) also increases with the square root of exposure time because noise obeys normal distribution. Therefore, the SNR (LOD) improves with the square root of exposure time.
The Raman scattering intensity of CO2 (1285 cm−1 and 1388 cm−1) acquired from single and double arms is shown in Table 1, the scattering intensity of carbon dioxide is obviously improved by double-arm collection.
For the quantitative analysis of CO2 concentration, the peak height of 1388 cm−1 is used to reflect the partial pressure of CO2. To acquire serial partial pressures of CO2 over a wide range, three standard gases with different CO2 concentrations (5000 ppm, 500 ppm, 50 ppm, carried by Ar) are used to fill the gas cell at different total pressures. The results shown in Fig. 9 demonstrate that the peak height of CO2 has an excellent linear relationship with its partial pressure.
3.3. Acetylene
The Raman spectrum of 100 kPa C2H2 standard gas (5000 ppm C2H2 carried by Ar, the total pressure in the cell is 104 kPa) with an exposure time of 200 s and a slit width of 100 µm is shown in Fig. 10(a). The spectrum is mainly composed of 12C2H2 (including the ν1 transition observed at 3372 cm−1, the ν2 transition observed at 1972 cm−1, the hot bands observed at 1960 cm−1 and 3357 cm−1), 12C13CH2 (including the ν1 transition observed at 1941 cm−1 and the ν2 transition observed at 3361 cm−1), and the remained N2. O- and S-branches of C2H2 can also be clearly demonstrated after zooming in the spectrum, as shown in Figs. 10(b) and 10(c).
Fig. 10. The Raman spectrums of C2H2. (a) An overview Raman spectrum of C2H2, including 12C2H2, 12C13CH2, and remained N2. (b) O- and S-branches of ν1 transition of 12C2H2, ν1 transition of 12C13CH2 also observed at 1941 cm−1. (c) O- and S-branches of ν2 transition of 12C2H2, Q-branch of ν2 transition of 12C13CH2 also observed at 3361 cm−1.
With 200 s exposure time and a slit width of 100 µm, a SNR of 2600 and a corresponding LOD of 0.58 Pa for C2H2 (1972 cm−1) are achieved. For the Q-branch of the remained N2, a SNR of 1750 and a corresponding LOD of 5.35 Pa are achieved. The intensity of Raman scattering of C2H2 acquired from single and double arms is shown in Table 1, the scattering intensity of C2H2 is obviously improved by double-arm collection.
For quantitative analysis, the peak height of 1972 cm−1 (ν1 transition of 12C2H2) is used to reflect the partial pressure of C2H2. To acquire serial partial pressures of C2H2 over a wide range, three standard gases with different C2H2 concentration (5000 ppm, 500 ppm, 50 ppm, carried by Ar) are used to fill the gas cell at different total pressures. For C2H2, the peak height also has an excellent linear relationship with the partial pressure, as shown in Fig. 11. It also indicates that the accurate quantitative analysis of trace gases is capable based on CERS.
4. Conclusion
In this paper, we report a cavity-enhanced Raman spectroscopy (CERS) apparatus for gas sensing. With the optical feedback frequency-locking, the laser can be effectively coupled to the enhanced-cavity. The direct reflection light does not return to the laser source with the help of the three-mirror V-shaped cavity configuration. Using a 42 mW diode laser, a 92 W intracavity laser power can be built up, yielding a power gain factor of 2200. Raman signals output from the two cavity-arms are collected simultaneously in a forward scattering collection geometry based on the polarization of scattered light. Raman spectrums of air, carbon dioxide, and acetylene are recorded to demonstrate the performance of the CERS gas-sensing apparatus. Multicomponent gas mixtures including their isotopic gases can be simultaneously sensed by CERS. With an exposure time of 200 s, LODs of 5.35 Pa for N2, 5.07 Pa for O2, 1.74 Pa for CO2, and 0.58 Pa for C2H2 are achieved. At 1 bar total pressure, this corresponds to 53.5, 50.7, 17.4, and 5.8 ppm by volume. CERS has been verified with high selectivity and sensitivity for gas sensing in this paper. Moreover, the intensity of Raman signals (peak height) has an excellent linear relationship with the partial pressure of the detected gases, which means CERS also has the potential for quantitative analysis of gases with high accuracy.
Funding
National Natural Science Foundation of China (61605020, U1766217).
References
1. L. B. Kreuzer, “Ultralow gas concentration infrared absorption spectroscopy,” J. Appl. Phys. 42(7), 2934–2943 (1971). [CrossRef]
2. K. A. Tillman, R. R. J. Maier, D. T. Reid, and E. D. Mcnaghten, “Mid-infrared absorption spectroscopy of methane across a 14.4THz spectral range using a broadband femtosecond optical parametric oscillator based on aperiodically poled lithium niobate,” J. Opt. A: Pure Appl. Opt. 7(6), S408–S414 (2005). [CrossRef]
3. S. W. Sharpe, R. Sheeks, C. Wittig, and R. A. Beaudet, “Infrared absorption spectroscopy of CO2-Ar complexes,” Chem. Phys. Lett. 151(3), 267–272 (1988). [CrossRef]
4. 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 monitoring,” Nat. Commun. 8(1), 15331 (2017). [CrossRef]
5. L. Dong, A. A. Kosterev, D. Thomazy, and F. K. Tittel, “QEPAS spectrophones: design, optimization, and performance,” Appl. Phys. B: Lasers Opt. 100(3), 627–635 (2010). [CrossRef]
6. H. P. Wu, L. Dong, X. K. Yin, A. Sampaolo, P. Patimisco, W. G. Ma, L. Zhang, W. B. Yin, L. T. Xiao, V. Spagnolo, and A. S. T. Jia, “Atmospheric CH4 measurement near a landfill using an ICL-based QEPAS sensor with V-T relaxation self-calibration,” Sens. Actuators, B 297, 126753 (2019). [CrossRef]
7. S. Z. Li, L. Dong, H. P. Wu, A. Sampaolo, P. Patimisco, V. Spagnolo, and 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. J. Peltola, M. Vainio, T. Hieta, J. Uotila, S. Sinisalo, M. Metsala, M. Siltanen, and L. Halonen, “High sensitivity trace gas detection by cantilever-enhanced photoacoustic spectroscopy using a mid-infrared continuous-wave optical parametric oscillator,” Opt. Express 21(8), 10240–10250 (2013). [CrossRef]
9. V. Spagnolo, L. Dong, A. A. Kosterev, and F. K. Tittel, “Modulation cancellation method for isotope 18O/16O ratio measurements in water,” Opt. Express 20(4), 3401–3407 (2012). [CrossRef]
10. C. V. Raman and K. S. Krishnan, “A new type of secondary radiation,” Nature 121(3048), 501–502 (1928). [CrossRef]
11. M. A. Buldakov, V. A. Korolkov, I. I. Matrosov, D. V. Petrov, A. A. Tikhomirov, and B. V. Korolev, “Analyzing natural gas by spontaneous Raman scattering spectroscopy,” J. Opt. Technol. 80(7), 426–430 (2013). [CrossRef]
12. W. R. Fenner, H. A. Hyatt, J. M. Kellam, and S. P. S. Porto, “Raman cross section of some simple gases,” J. Opt. Soc. Am. 63(1), 73–77 (1973). [CrossRef]
13. W. K. Bischel and G. Black, “Wavelength dependence of Raman scattering cross sections from 200-600 nm,” AIP Conf. Proc. 100(1), 181–187 (1983). [CrossRef]
14. R. V. Pound, “Electronic frequency stabilization of microwave oscillators,” Rev. Sci. Instrum. 17(11), 490–505 (1946). [CrossRef]
15. E. D. Black, “An introduction to Pound–Drever–Hall laser frequency stabilization,” Am. J. Phys. 69(1), 79–87 (2001). [CrossRef]
16. D. J. Taylor, M. Glugla, and R. D. Penzhorn, “Enhanced Raman sensitivity using an actively stabilized external resonator,” Rev. Sci. Instrum. 72(4), 1970–1976 (2001). [CrossRef]
17. V. Sandfort, J. Goldschmidt, J. Wollenstein, and S. Palzer, “Cavity-enhanced Raman spectroscopy for food chain management,” Sensors 18(3), 709 (2018). [CrossRef]
18. Y. F. Pan, L. Dong, H. P. Wu, W. G. Ma, L. Zhang, W. B. Yin, L. T. Xiao, S. T. Jia, and And F. K. Tittel, “Cavity-enhanced photoacoustic sensor based on a whispering-gallery-mode diode laser,” Atmos. Meas. Tech. 12(3), 1905–1911 (2019). [CrossRef]
19. A. J. Friss, C. M. Limbach, and A. P. Yalin, “Cavity-enhanced rotational Raman scattering in gases using a 20 mW near-infrared fiber laser,” Opt. Lett. 41(14), 3193–3196 (2016). [CrossRef]
20. R. Salter, J. Chu, and M. Hippler, “Cavity-enhanced Raman spectroscopy with optical feedback cw diode lasers for gas phase analysis and spectroscopy,” Analyst 137(20), 4669–4676 (2012). [CrossRef]
21. R. Keiner, T. Frosch, T. Massad, S. Trumbore, and J. Popp, “Enhanced Raman multigas sensing - a novel tool for control and analysis of (13)CO(2) labeling experiments in environmental research,” Analyst 139(16), 3879–3884 (2014). [CrossRef]
22. T. Frosch, R. Keiner, B. Michalzik, B. Fischer, and J. Popp, “Investigation of Gas Exchange Processes in Peat Bog Ecosystems by Means of Innovative Raman Gas Spectroscopy,” Anal. Chem. 85(3), 1295–1299 (2013). [CrossRef]
23. A. Sieburg, T. Jochum, S. E. Trumbore, J. Popp, and T. Frosch, “Onsite cavity enhanced Raman spectrometry for the investigation of gas exchange processes in the Earth’s critical zone,” Analyst 142(18), 3360–3369 (2017). [CrossRef]
24. 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]
25. T. Jochum, B. Michalzik, A. Bachmann, J. Popp, and T. Frosch, “Microbial respiration and natural attenuation of benzene contaminated soils investigated by cavity enhanced Raman multi-gas spectroscopy,” Analyst 140(9), 3143–3149 (2015). [CrossRef]
26. R. Keiner, T. Frosch, S. Hanf, A. Rusznyak, D. M. Hkob, K. Kusel, and J. Popp, “Raman spectroscopy—an innovative and versatile tool to follow the respirational activity and carbonate biomineralization of important cave bacteria,” Anal. Chem. 85(18), 8708–8714 (2013). [CrossRef]
27. S. Hanf, S. Fischer, H. Hartmann, R. Keiner, S. Trumbore, J. Popp, and T. Frosch, “Online investigation of respiratory quotients in Pinus sylvestris and Picea abies during drought and shading by means of cavity-enhanced Raman multi-gas spectrometry,” Analyst 140(13), 4473–4481 (2015). [CrossRef]
28. R. Keiner, M. C. Gruselle, B. Michalzik, J. Popp, and T. Frosch, “Raman spectroscopic investigation of (CO2)-C-13 labeling and leaf dark respiration of Fagus sylvatica L. (European beech),” Anal. Bioanal. Chem. 407(7), 1813–1817 (2015). [CrossRef]
29. D. A. King and R. J. Pittaro, “Simple diode pumping of a power-buildup cavity,” Opt. Lett. 23(10), 774–776 (1998). [CrossRef]
30. J. Morville, D. Romanini, A. A. Kachanov, and M. Chenevier, “Two schemes for trace detection using cavity ringdown spectroscopy,” Appl. Phys. B: Lasers Opt. 78(3-4), 465–476 (2004). [CrossRef]
31. S. Ohshima and H. Schnatz, “Optimization of injection current and feedback phase of an optically self-locked laser diode,” J. Appl. Phys. 71(7), 3114–3117 (1992). [CrossRef]
32. S. G. Baran, G. Hancock, R. Peverall, G. A. D. Ritchie, and N. J. Van Leeuwen, “Optical feedback cavity enhanced absorption spectroscopy with diode lasers,” Analyst 134(2), 243–249 (2009). [CrossRef]
33. P. Spano, S. Piazzolla, and M. Tamburrini, “Theory of noise in semiconductor lasers in the presence of optical feedback,” IEEE J. Quantum Electron. 20(4), 350–357 (1984). [CrossRef]
34. G. P. Agrawal and C. H. Henry, “Modulation performance of a semiconductor laser coupled to an external high-Q resonator,” IEEE J. Quantum Electron. 24(2), 134–142 (1988). [CrossRef]
35. H. Li and N. B. Abraham, “Analysis of the noise spectra of a laser diode with optical feedback from a high-finesse resonator,” IEEE J. Quantum Electron. 25(8), 1782–1793 (1989). [CrossRef]
36. J. Morville, S. Kassi, M. Chenevier, and D. Romanini, “Fast, low-noise, mode-by-mode, cavity-enhanced absorption spectroscopy by diode-laser self-locking,” Appl. Phys. B: Lasers Opt. 80(8), 1027–1038 (2005). [CrossRef]
37. R. Lang and K. Kobayashi, “External optical feedback effects on semiconductor injection laser properties,” IEEE J. Quantum Electron. 16(3), 347–355 (1980). [CrossRef]
38. R. Lang, “Injection locking properties of a semiconductor laser,” IEEE J. Quantum Electron. 18(6), 976–983 (1982). [CrossRef]
39. E. M. Purcell, “Spontaneous emission probabilities at radio frequencies,” Phys. Rev. 69(11-1), 681 (1946).
40. K. J. Vahala, “Optical microcavities,” Nature 424(6950), 839–846 (2003). [CrossRef]
41. Y. Wu, X. X. Yang, and P. T. Leung, “Theory of microcavity-enhanced Raman gain,” Opt. Lett. 24(5), 345–347 (1999). [CrossRef]
42. T. Hummer, J. Noe, M. S. Hofmann, T. W. Hansch, A. Hogele, and D. Hunger, “Cavity-enhanced Raman microscopy of individual carbon nanotubes,” Nat. Commun. 7(1), 12155 (2016). [CrossRef]
43. B. Petrak, N. Djeu, and A. Muller, “Purcell enhancement of Raman scattering from atmospheric gases in a high-finesse microcavity,” Phys. Rev. A 89(2), 023811 (2014). [CrossRef]
44. N. Ismail, C. C. Kores, D. Geskus, and M. pollnau, “Fabry-Perot resonator: spectral line shapes, generic and related Airy distributions, linewidths, finesses, and performance at low or frequency-dependent reflectivity,” Opt. Express 24(15), 16366–16389 (2016). [CrossRef]
45. J. Bendtsen, “High-resolution Fourier transform Raman spectra of the fundamental bands of 14N15N and 15N2,” J. Raman Spectrosc. 32(12), 989–995 (2001). [CrossRef]
46. Trace gas detection of molecular hydrogen H2 by photoacoustic stimulated Raman spectroscopy (PARS)
47. B. Petrak, J. Copper, K. Konthasinghe, M. Peiris, N. Djeu, A. J. Hopkins, and A. Muller, “Isotopic gas analysis through Purcell cavity enhanced Raman scattering,” Appl. Phys. Lett. 108(9), 091107 (2016). [CrossRef]
