Open Access
Add to BibSonomyAbstract
We report on sensitive detection of atmospheric methane employing quantum cascade laser based optical feedback cavity-enhanced absorption spectroscopy (OF-CEAS). An instrument has been built utilizing a continuous-wave distributed feedback quantum cascade laser (cw-QCL) with a V-shaped cavity, a common arrangement that reduces feedback to the laser from non-resonant reflections. The spectrometer has a noise equivalent absorption coefficient of 3.6 × 10−9 cm−1 Hz−1/2 for a spectral scan of CH4 at 7.39 μm. From an Allan-Werle analysis a detection limit of 39 parts per trillion of CH4 at atmospheric pressure within 50 s acquisition time was found.
© 2016 Optical Society of America
1. Introduction
The importance of sensitive trace gas detection is manifold. Not only in industrial process control and for the monitoring of air pollutants and greenhouse gases, but also in medical breath analysis and for the detection of drugs and explosives a sensitive, real-time and selective detection of trace gases is needed. Here, the focus will be on the measurement of the greenhouse gas methane CH4, which is the second most important anthropogenic greenhouse gas after CO2. CH4 has grown from a pre-industrial average atmospheric mole fraction of about 722 parts per billion (pbb) to 1833±1 ppb in 2014, an increase of 9 ppb with respect to the previous year [1]. There is also a pronounced seasonal cycle with on average 25 ppb less methane in the atmosphere of the northern hemisphere during summer when compared with the rest of the year. Although its atmospheric abundance is much lower than for CO2 (397.7 parts per million (ppm)), it has an estimated global warming potential per molecule which is ~34 times greater than for CO2 over a 100 year horizon and its increasing abundance in the atmosphere could lead to an acceleration of the ongoing climate change. For this reason, accurate and sensitive measurements of its concentration have become of central importance in current research. Ground-based observations of greenhouse gases are not only useful for obtaining local information but are also often indispensable for the validation of satellite measurements. In the meantime various optical techniques for the quantification of atmospheric trace gases have been developed, such as differential optical absorption spectroscopy, LIDAR and cavity-enhanced spectroscopy (CES) techniques like cavity ring-down spectroscopy (CRDS) and cavity-enhanced absorption spectroscopy (CEAS). Especially, CES techniques, such as off-axis CEAS, optical feedback CEAS and noise-immune cavity-enhanced optical heterodyne molecular spectroscopy (NICE-OHMS), have been shown to be very sensitive. For more information on the developments and applications of CES techniques, we refer the reader to two recent scientific books on this topic and the references therein [2, 3]. It has among others been demonstrated that a compact and robust spectrometer based on OF-CEAS is ideally suited for field measurements, for example aboard aircrafts [4] or for exploring past climate by embedding the spectrometer directly in the drilling probe untilized for ice drilling in Antarctica [5]. As the strongest absorptions of many molecules of both fundamental and applied interest, including CH4, lie in the mid-infrared spectral region (MIR), OF-CEAS based spectrometers operating in this spectral region provide an exciting prospect for further improvement of the detection limit, see [6–10]. In 2011, Hamilton et al. reported first OF-CEAS measurements on CH4 using a continuous-wave distributed feedback quantum cascade laser (cw-QCL) emitting at 7.84 μm with an output power of up to 4 mW [7]. With a bandwidth reduced sensitivity of 5.5 × 10−8 cm−1 Hz−1/2 (1σ) they found a detection limit of 8 ppb of CH4 in air at atmospheric pressure. We recently reported on the use of a continuous-wave distributed feedback interband cascade laser (cw-ICL) operating at ambient temperature as the light source in a OF-CEAS experiment to detect CH4 in the 3 μm spectral region [10, 11]. Although a detection limit of 3 ppb CH4 at atmospheric pressure was achieved, comparable to previously reported OF-CEAS instruments, the low output power of the cw-ICL was a drawback in achieving ultimate sensitivity. Therefore, we now have chosen the CH4 absorption band around 7.37 μm as our target in view of the good availability of a high power cw-QCL, while we only forfeit a factor of 4 in the linestrength of the transitions measured. In this paper, we demonstrate a detection limit for CH4 down to the part per trillion level (ppt) employing OF-CEAS utilizing a powerful cw-QCL source with a V-shaped cavity, a commonly used arrangement that reduces feedback to the laser from nonresonant reflections.
2. Experimental details
The cw-QCL (AdTech optics) employed for the OF-CEAS measurements was housed in an industrial standard high heat load package (HHL) with collimation lens enclosed and mounted on a custom-made water-cooled heat sink. The cw-QCL delivers an output power of up to 180 mW at 7.37 μm. Temperature was thermoelectrically controlled by a PTC10k-CH driver and current supplied by a QCL1000 driver (both Wavelength Electronics). The laser was scanned ~0.55 cm−1 by applying a sawtooth function at 20 Hz produced by a data acquisition card (DAQ; National Instruments PCI-6221, 16 bit resolution, 740 kS/s), resulting in an acquisition time of 50 ms per scan. Particular care has been taken to operate the cw-QCL as stable as possible to ensure maximum performance of the spectrometer. To prevent interference from additional digital noise originating from the DAQ card a low-pass filter with a 3 dB frequency of 663 kHz was implemented before the input of the current driver. With the help of an additional voltage divider between DAQ card and current driver the maximum digital resolution of the DAQ card could be utilized for driving the cw-QCL. All data were collected with the laser scanning to lower wavenumber. As shown in Fig. 1, the MIR beam was directed into a sealed aluminum box with one steering mirror mounted on a piezoelectric transducer (PZT; Physik Instrumente) for fine control of the feedback phase. The BaF2 windows were tilted with respect to the optical axis to reduce reflections returning to the laser. Inside the box, a V-shaped cavity with each cavity arm length being 40 cm was constructed from three spherical mirrors M0, M1, and M2 to give a free spectral range (FSR) and therefore a spectral sampling resolution of 187.5 MHz. The mirrors (LohnStar) have a specified high reflectivity coating of R ≤ 0.99985 with a reflectivity peak at 7.4 μm on a 1 inch diameter ZnSe substrate with a 1 m radius of curvature. The beam was injected into the cavity at the folding mirror (M0) with the laser-to-cavity distance set equal to each cavity arm length. The optical feedback rate to the cw-QCL was controlled with the help of an aperture stop (AS) in front of the steering mirror. For monitoring of the laser intensity before cavity injection and determination of the free-running scan rate an additional beam splitter (BS) was inserted realizing a by-pass. An off-axis parabolic mirror (OAPM) (1 inch focal length) outside the box focused the light leaking through the first cavity mirror (M1) onto a DC-coupled photovoltaic detector (neoplas control IRDM-DCA-10.6). Using a flip mirror (FM) at the exit of the cavity the same detector could be used for all measurements. The transmitted signal was recorded on a 25 MHz bandwidth DAQ card (Alazar Tech ATS 300, 12 bit, 50 MS/s). Further averaging and processing of the data was performed using custom LabVIEW routines on a Mini-ITX computer (ADLINK Technology). The Lab-VIEW program generated an error signal for the PZT controller (Physik Instrumente E-505.00) based on the symmetry of the cavity modes for each recorded scan and the PZT for maintaining phase-locking was controlled via the same DAQ card as used for scanning the laser. In the experiment, the laser was scanned across three strong methane transitions around 1353 cm−1. The data reported here were measured with nominal mixing ratios of 300 ppb and 500 ppb CH4 in N2 as buffer gas. The samples were formed by buffering a commercial gas mixture of 10 ppm CH4 diluted in N2 with additional N2 (both Air Liquide). In order to obtain a pressure broadened absorption lineshape covering a sufficient number of cavity FSRs for spectral analysis, spectra of both samples were collected at various pressures from 15 to 50 kPa. Pressure was measured by a capacitive transducer (Oerlikon Leybold Vacuum).
Fig. 1 Schematic of the OF-CEAS experimental setup. The shaded box indicates the sealed sample container and the dotted line shows the phase-locking control loop. Note: M0 – M2 = spherical mirrors, PVD = photovoltaic detector, OAPM = off-axis parabolic mirror, FM = flip mirror, PZT = piezoelectric transducer, AS = aperture stop, BS = beam splitter.
3. Results and discussion
In Fig. 2(a), a typical cavity transmission spectrum averaged over 1000 scans of a 30 kPa sample with 300 ppb CH4 in N2 as buffer gas is presented. Over ~45 ms, the cw-QCL locked to 97 successive cavity modes with a locking time of ~430 μs, much longer than the ring-down time (typically 9.2 μs). The time between modes was ~220 μs (Fig. 2(b)). Besides utilizing the aperture stop in front of the cavity to control the feedback rate and thereby ensuring the frequency locking range to be smaller than on cavity FSR also the time between modes could be made large enough for the determination of the detector offset. As the non-zero signal between the modes is considered as the detector offset, it has been determined and subtracted by a linear fit to the signal between the locking modes. No broadening of the modes was observed due to averaging of the spectra. There is a clear amplitude oscillation in the transmitted signal for alternating modes due to even and odd modes having a different phase at the folding mirror, and therefore slightly different reflectivity. The upper traces in Fig. 2 were measured by using a Ge etalon in the by-pass. Clearly horizontal steps can be seen in the etalon signal, an explicit indication of optical feedback locking of the laser frequency to the modes of the cavity. For this experiment, stable feedback locking was achieved without applying mode matching as has often been done in past OF-CEAS studies. The power of the laser behind the aperture stop was 21 mW. Using the back reflection from the folding mirror it was found that ~17% of the power was injected into the cavity in resonance. The power on the detector behind the cavity was measured to be about 280 μW showing that almost 9% of the injected power is transmitted at the mode maxima. Furthermore, it was found that the effect of a reduction of optical feedback due to intracavity absorption was negligible. The measured laser power variation was less than 1% (see etalon trace in Fig. 2(b)) and therefore no correction was made for this effect.
Fig. 2 a) Cavity transmission for 300 ppb CH4 in N2 as buffer gas at a total pressure of 30 kPa while the laser is scanning to lower wavenumber. This transmission spectrum is the average of 1000 scans. b) A magnification of the modes around 45 ms. The locking time of the modes was ~430 μs and the time between modes was ~220 μs. In both figures, the upper trace is an etalon trace showing flat sections when locking to the modes occurs (shifted for clarity).
For the determination of the absorption coefficient α the well-known formula was used assuming that sample absorption is small compared to mirror losses [7, 12]:
with I being the amplitude of each cavity mode and I0 being the amplitude of the cavity modes representing the baseline, which was determined by a 2nd order polynomial fit of the amplitudes of both the first and the last 8 modes of the scan. The property L is the physical length of the cavity of 80 cm. The effective reflectivity R was determined by measuring the integrated absorption for several CH4 samples with known concentration. To that end, samples of 500 ppb CH4 in N2 as buffer gas were analyzed in the pressure range of 15 to 50 kPa. The absorption spectrum at 1353 cm−1 consists primarily of three transitions in the R branch of the triply degenerate ν4 bending mode of CH4: 9 A2 1 ← 8 A1 1 (ν = 1353.159 cm−1); 9 F2 2 ← 8 F1 1 (ν = 1353.075 cm−1); 9 E 2 ← 8 E 1 (ν = 1353.026 cm−1). The assignments were taken from the HITRAN database [13]. For the analysis, the more than two orders of magnitude weaker transition at 1353.097 cm−1 was also taken into account. In Fig. 3, the integrated CEAS signal measured as a function of the CH4 concentration in N2 as buffer gas at different total pressures is shown. The analysis had to be performed separately for even and odd modes to account for the different phase at the folding mirror. For both cases a linear relationship was found. With the effective line strength of 8.59×10−20 cm2 cm−1 molec−1 [13] for the four transitions in the spectrum the mean value for R could be determined from the slope for both even and odd modes, which were found to be 0.9997 and 0.99971, respectively with a common standard variation of ±0.00001. This corresponds, using the mean reflectivity, to an optical cavity finesse of πR/(1 − R2) = 5416 and correspondingly an effective path length of 2760 m.Fig. 3 The integrated CEAS signal as a function of the CH4 concentration, 500 ppb in N2 as buffer gas, at different total pressures, is shown with a linear fit through the data points. For even and odd modes the slope gives a value for R of 0.9997 and 0.99971, respectively.
Figure 4 shows a cumulative fit to the measured absorption coefficient spectrum of 300 ppb CH4 in N2 as buffer gas at a total pressure of 30 kPa averaged over 1000 scans. The baseline intensity I0 was here determined via a 2nd order polynomial fit of the amplitudes of both the first and last 12 modes of the scan. In the lower panel of Fig. 4, the residual of the fitted absorption profile is shown. The standard deviation of the residual of the cumulative fit gives a minimum detectable absorption coefficient, αmin, of 3.4 × 10−8 cm−1 (1σ). From this value, using the relation αmin = σpeaknmin, a detection limit of 2.5 ppb CH4 at atmospheric pressure can be derived. The peak absorption cross section, σpeak, is calculated from the line strength taking into account pressure broadening of the line shape using the pressure broadening coefficient, both given in [13]. It is obvious from both the spectrum and the residual shown in Fig. 4 that the limit of detection of 2.5 ppb is hampered by a sinusoidal structure in the baseline. This is a well-known effect often observed in CRDS and CEAS, due to optical interference with parasitic reflections in the setup [14, 15].
Fig. 4 Upper panel: Absorption coefficient spectrum of 300 ppb of CH4 in N2 as a buffer gas at a total pressure of 30 kPa together with a composite fit taking into account the 4 transitions of CH4 which contribute to the observed absorption feature. Lower panel: the residual of the fitted absorption profile; the standard deviation of the residual of this fit gives αmin = 3.35 × 10−8 cm−1.
To characterize the ultimate sensitivity as well as the temporal stability of the reported OF-CEAS setup, a type of Allan-Werle analysis was performed in which the minimum detectable absorption coefficient, αmin, was measured as a function of the acquisition time τ (number of scans × 50 ms). According to common practice in the characterization of cavities αmin was derived by analyzing the noise of the spectrum in case of zero absorption, i.e., in absence of any absorption feature [8, 16, 17]. Furthermore, Gorrotxategi-Carbajo and co-workers even claim that to estimate a detection limit for a given species a measurement with zero concentration of that species should be performed, because possible variations in non-zero concentrations can cause a rise of excess noise leading to an overestimation of the detection limit [8]. Therefore, the cavity was filled with pure N2 at a pressure of 30 kPa. In order to overcome the problem of the sinusoidal structure in the baseline, a background spectrum I0 was recorded for the same number of scans and used for the calculation of α according to Eq. 1 instead of fitting a polynomial to the baseline. In this way, the effect of the sinusoidal structure in the spectrum could be significantly reduced. Using the standard deviation of α as the minimum detectable absorption coefficient, the resulting αmin was tracked and recorded over extensive time periods for a particular acquisition time. The mean for αmin value was plotted against the acquisition time τ whereas its standard deviation serves as an uncertainty for this value.
In Fig. 5, the result of this type of Allan-Werle analysis performed with and without running the vacuum pump connected to the setup is shown. In both cases similar sensitivities for all acquisitions times were found. The Allan-Werle deviation for acquisition times shorter than 10 s decreases with a slope of −0.5 as expected for white noise in the system. The increasing deviation for long integration times is the result of system drifts. The best value of αmin of (5.3 ± 0.1) × 10−10 cm−1 was obtained for 1000 averages recorded in 50 s. This corresponds to a minimum detectable concentration of 39 ppt of CH4 at atmospheric pressure, which is the smallest value reported for CH4 concentration measurements using OF-CEAS. The bandwidth normalized sensitivity can be specified as 3.7×10−9 cm−1 Hz−1/2 which corresponds to 276 ppt CH4 at atmospheric pressure in 1 s acquisition time.
Fig. 5 Allan deviation plot of the minimum detectable sensitivity, αmin, derived from the comparison of a stored background to the cavity transmission for each mode, as a function of acquisition time τ. The cavity was filled with pure nitrogen at a pressure of 30 kPa. Blue circles correspond to data collected with a running pump in the spectrometer while black squares collected with this pump disabled.
4. Conclusion
It was demonstrated that using optical feedback cavity-enhanced absorption spectroscopy with quantum cascade lasers is an appropriate method for very sensitive detection of methane. Choosing the CH4 absorption band at 7.37μm a high power cw-QCL could be successfully employed in a standard V-shaped cavity without any mode matching. With a tuning frequency of 20 Hz the laser locked to more than 90 successive cavity modes covering a spectral range of 0.55 cm−1. To prevent frequency locking of individual modes larger than one cavity FSR the optical feedback rate had to be reduced with the help of an aperture stop. The resulting locking time was ~430 μs whereas the time between the modes was ~220 μs. Analyzing a CH4 absorption at 300 ppb, from the residuals of the fit an αmin of 3.35 × 10−8cm−1 (1σ) was found, which corresponds to a detection limit for CH4 of 2.5 ppb at atmospheric pressure. This detection limit is already sufficient to measure typically changes in the CH4 abundances in the atmosphere. Finally, to characterize the ultimate sensitivity of the reported OF-CEAS setup a type of Allan-Werle analysis was performed in an absorber free cavity at 30 kPa. A minimum detectable absorption coefficient of (5.3 ± 0.1) × 10−10 cm−1 (1σ) was found with a 50 s acquisition time. This corresponds to a CH4 concentration of 39 ppt at atmospheric pressure. The achieved bandwidth normalized sensitivity was 3.6 × 10−9 cm−1 Hz−1/2 which corresponds to 276 ppt of CH4 at atmospheric pressure in 1 s acquisition time. There is significant and real potential that these limits can be further extended with reasonably attainable improvements to the experimental system. The current sensitivity was achieved with a cavity of relatively low finesse, which could be improved by the use of higher-reflectivity mirrors.
Acknowledgments
This work was supported by the German Federal Ministry of Education and Research Grant VIP, FKZ 03V0122. The authors thank F. Weichbrodt for his dedicated and valuable technical support.
References and links
1. World Metrology Organization, “The state of greenhouse gases in the atmosphere based on global observations through 2014,” WMO-GAW Greenhouse Gas Bulletin11 (2015).
2. G. Berden and R. Engeln, eds. Cavity Ring-Down Spectroscopy: Techniques and Applications (Wiley-Blackwell, 2009).
3. G. Gagliardi and H.-P. Loock, eds. Springer Series in Optical Sciences Vol. 179: Cavity-Enhanced Spectroscopy and Sensing (Springer, 2014). [CrossRef]
4. D. Romanini, M. Chenevier, S. Kassi, M. Schmidt, C. Valant, M. Ramonet, J. Lopez, and H.-J. Jost, “Optical-feedback cavity-enhanced absorption: a compact spectrometer for real-time measurement of atmospheric methane,” Appl. Phys. B 83, 659–667 (2006). [CrossRef]
5. R. Grilli, N. Marrocco, T. Desbois, C. Guillerm, J. Triest, E. Kerstel, and D. Romanini, “SUBGLACIOR: An optical analyzor embedded in an antartic ice probe for exploring the past climate,” Rev. Sci. Instrum. 85, 111301 (2014). [CrossRef]
6. G. Maisons, P. Gorrotxategi-Carbajo, M. Carras, and D. Romanini, “Optical-feedback cavity-enhanced absorption spectroscopy with a quantum cascade laser,” Opt. Lett. 2010, 3607–3609 (2010). [CrossRef] [PubMed]
7. D. J. Hamilton and A. J. Orr-Ewing, “A quantum cascade laser-based optical feedback cavity-enhanced absorption spectrometer for the simultaneous measurement of CH4 and N2O in air,” Appl. Phys. B 102, 879–890 (2011). [CrossRef]
8. P. Gorrotxategi-Carbajo, E. Fasci, I. Ventrillard, M. Carras, G. Maisons, and D. Romanini, “Optical-feedback cavity-enhanced absoprtion spectroscopy with a quantum-cascade laser yields the lowest formaldehyde detection limit,” Appl. Phys. B 110, 309–314 (2013). [CrossRef]
9. A. G. V. Bergin, G. Hancock, G. A. D. Ritchie, and D. Weidmann, “Linear cavity optical-feedback cavity-enhanced absorption spectroscopy with a quantum cascade laser,” Opt. Lett. 38, 2475–2477 (2013). [CrossRef] [PubMed]
10. K. M. Manfred, G. A. D. Ritchie, N. Lang, J. Röpcke, and J. H. van Helden, “Optical feedback cavity-enhanced absorption spectroscopy with a 3.24 μ m interband cascade laser,” Appl. Phys. Lett. 106, 221106 (2015). [CrossRef]
11. J. H. van Helden, N. Lang, U. Macherius, H. Zimmermann, S. Glitsch, M. Wiese, and J. Röpcke, “On Recent Progress Applying Quantum Cascade Lasers in Sensing for Environmental and Plasma Diagnostics,” in Light, Energy and Environment 2015, OSA Technical Digest (Optical Society of America, 2015), paper ETh2A.1. [CrossRef]
12. 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 80, 1027–1038 (2005). [CrossRef]
13. L. S. Rothman, I. E. Gordon, Y. Babikov, A. Barbe, D. C. Benner, P. F. Bernath, M. Birk, L. Bizzocchi, V. Boudon, L. R. Brown, A. Campargue, K. Chance, E. A. Cohen, L. H. Coudert, V. M. Devi, B. J. Drouin, A. Fayt, J.-M. Flaud, R. R. Gamache, J. J. Harrison, J.-M. Hartmann, C. Hill, J. T. Hodges, D. Jacquemart, A. Jolly, J. Lamouroux, R. J. LeRoy, G. Li, D. A. Long, O. M. Lyulin, C. J. Mackie, S. T. Massie, S. Mikhailenko, H. S. P. Müller, O. V. Naumenko, A. V. Nikitin, J. Orphal, V. Perevalov, A. Perrin, E. R. Polovtseva, C. Richard, M. A. H. Smith, E. Starikova, K. Sung, S. Tashkun, J. Tennyson, G. C. Toon, V. G. Tyuterev, and G. Wagner, “The HITRAN2012 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transfer 130, 4–50 (2012). [CrossRef]
14. S. Kassi, M. Chenevier, L. Gianfrani, A. Salhi, Y. Rouillard, A. Ouvrard, and D. Romanini, “Looking into the volcano with a MIR-IR DFB diode laser and Cavity Enhanced Absorption Spectroscopy,” 14, 11442–11452 (2006).
15. C. L. Bell, J. P. H. van Helden, T. P. J. Blaikie, G. Hancock, N. J. van Leeuwen, R. Peverall, and G. A. D. Ritchie, “Noise-immune cavity-enhanced optical heterodyne detection of HO2 in the near-infrared range,” J. Phys. Cem. A 116, 5090–5099 (2012).
16. L. Menzel, A. A. Kosterev, R. F. Curl, F. K. Tittel, C. Gmachl, F. Capasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho, and W. Urban, “Spectroscopic detection of biological NO with a quantum cascade laser,” Appl. Phys. B 72, 859–863 (2001). [CrossRef]
17. I. Courtillot, J. Morville, V. Motto-Ros, and D. Romanini, “Sub-ppb NO2 detection by optical feedback cavity enhanced absorption spectroscopy with a blue diode laser,” Apl. Phys. B 85, 407–412 (2006). [CrossRef]
