Open Access
Add to BibSonomy
Abstract
A new heterodyne interferometric method for optical signal detection in photoacoustic or photothermal spectroscopy is demonstrated and characterized. It relies on using one laser beam for the photoacoustic excitation of the gas sample that creates refractive index changes along the beam path, while another laser beam is used to measure these changes. A heterodyne-based detection of path-length changes is presented that does not require the interferometer to be balanced or stabilized, which significantly simplifies the optical design. We discuss advantages of this new approach to photoacoustic signal detection and the new sensing arrangements that it enables. An open-path photoacoustic spectroscopy of carbon dioxide at 2003 nm and a novel sensing configuration that enables three-dimensional spatial gas distribution measurement are experimentally demonstrated.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Laser gas-phase spectroscopy is a powerful tool in chemical analysis with applications spanning from environmental monitoring, through medical diagnostics, to leak detection and industrial process control. Photo-acoustic spectroscopy (PAS) is one of the techniques that is often used for chemical detection through indirect measurement of sample absorption [1]. PAS relies on photoacoustic (PA) effect primarily produced via photothermal excitation of the sample in which absorption of light produces acoustic wave via sample heating. In PAS the radiation source is modulated at audio or ultra-sound frequencies causing the sample temperature to change periodically in time and generate pressure (sound) waves. This PA signal is usually further enhanced by implementing an acoustic resonator into the gas cell and can be detected using a sensitive microphone. Other transducers, e.g. quartz tuning forks, can also be used for PA signal detection [2]. With the development of quantum cascade lasers (QCLs) providing single frequency operation in the mid-infrared molecular fingerprint region, and output powers above 100 mW, compact instruments offering ultrasensitive detection of various molecules at ppm (parts per million) or ppb (parts per billion) levels have been routinely demonstrated using PAS or quartz-enhanced PAS (QEPAS) techniques [3–6].
Since PA waves exhibit periodic pressure variations, sample density and thus refractive index are also periodically modulated. An interesting and so far not fully explored approach to PA signal retrieval is based on all-optical detection of these refractive index changes. This method relies on probing the refractive index with one beam that is aligned collinearly with a second beam that is used for photoacoustic excitation of the gas sample. In [7] this type of detection was demonstrated, using a modulated CO2 laser at 9.22 µm to produce PA signal in ammonia and a HeNe laser at 633 nm to perform signal measurement. Visible light was divided into two beams with one of them being aligned collinearly with the infrared light (so that it was experiencing generated refractive index changes). The two beams were subsequently detected with two photodiodes and balanced detection was used to retrieve phase changes due to PA effect. Similar concept was also explored in [8]. More recently, a number of photoacoustic (or photothermal) spectrometers with solely optical-based detection were demonstrated, usually with near-infrared excitation. These examples use either laser [9, 10] or broadband sources [11, 12] for PA signal generation and interferometric setup for its detection.
The most interesting property of optical-based detection in PAS is that it allows for measuring signals in any spatial location that can be accessed by the two beams, pump that generates PA signal and probe that is used to measure it. Therefore, various configurations are possible, including remote sensing or path integrated detection with sensing path extended using multi-pass cells or inside hollow core fibers [10, 11].
In this paper we introduce new heterodyne interferometric method for optical PAS detection, which is experimentally demonstrated and characterized. Advantages of this new approach to PA signal detection and new sensing configurations that it enables are discussed.
2. Heterodyne interferometric photoacoustic signal detection
The schematic diagram of the heterodyne interferometric detection in PAS is shown in Fig. 1. A pump laser is used to generate PA signal by tuning its wavelength into resonance with the absorption line of the target molecule. This laser (ideally a semiconductor device) needs to be intensity- or wavelength-modulated (by using e.g. external chopper or direct injection current modulation). The pump beam, after passing through the sensing section is blocked or deflected using an optical filter, dichroic mirror or polarizer. The second laser source is used as a probe laser. Its output beam is divided using a beam splitter into two branches. One beam (probe beam) is aligned collinearly with the pump beam. After passing through the sensing section it is combined with the second (reference) beam, that is frequency shifted by frequency Ω (using e.g. acousto-optical modulator, AOM). The two beams (probe and reference) are recombined using a second beam splitter and focused onto a photodetector. The photodetector, which is a square-law device, mixes the two optical waves and generates a heterodyne signal at Ω (AOM is typically modulated in a 10 to 100 MHz range). In this configuration any changes in the refractive index that occur in the sensing section (i.e. PA signal) will affect the optical phase, which is detected directly as phase of the heterodyne beatnote. Those changes can be retrieved using conventional frequency demodulation at the carrier frequency Ω, a process well known from FM radio.
Fig. 1 Schematic diagram of a heterodyne-based detection in PAS setup. BS – beam splitter, F – low-pass filter for the near-IR measurement beam, M – mirror.
This type of detection has several advantages. Wavelengths of the probe and the pump lasers can be chosen independently. For example, mid-infrared excitation can provide access to strong fundamental ro-vibrational transitions while detection of optical path-length variations associated with PA effects can be performed in the near-infrared region, where high-quality, relatively inexpensive detectors and sources are available. Because the useful signal is encoded in frequency/phase of the heterodyne beatnote, not in its amplitude, measured signal is immune to variations of optical power that reaches the detector. Moreover, detecting frequency of the heterodyne beatnote does not require the interferometer to be precisely balanced and stabilized (in contrast to standard setups used for optical-based PA detection which require the two arms of the interferometer to be stabilized for operation at quadrature [10, 13]). These properties give perspective for application of detection schemes that cannot be easily implemented with standard PAS, e.g. open-path sensing, PAS with long interaction length or ‘standoff PAS’ (Fig. 2).
Fig. 2 Schematic diagram of potential configurations of the ‘sensing section’ in Fig. 1 that allow for (a) standoff photoacoustic detection, or with optical pathlength enhancement using (b) multi-pass cell or (c) hollow core fiber.
3. Experiments
3.1 A proof-of-concept demonstration
Figure 3 shows a schematic diagram of the experimental layout used for initial experiments with hydrogen cyanide (HCN) as the target gas. As the pump source a distributed feed-back (DFB) laser diode at ~1548 nm was used. The wavelength of the pump laser was tuned to coincide with the P8 absorption line in the HCN 2v3 rotational-vibrational band at 1548.2 nm. A 15-cm long gas cell containing pure HCN at pressure of 10 Torr was used in the experiment. The pump laser output was amplified using an erbium/ytterbium doped fiber amplifier (EYDFA) providing optical power of approximately 200 mW. A mechanical chopper was used to modulate pump radiation at frequency fm = 600 Hz. The probe section was built as a fiber-based Mach-Zehnder heterodyne interferometer with a gas cell in one arm and a frequency shifter (AOM driven at Ω = 40 MHz) in the other. A DFB laser diode operating at ~1552 nm was used as a probe source with its wavelength adjusted such that it did not coincide with any HCN transition in this spectral region. A band pass filter (BPF) was used to attenuate remaining pump radiation. The heterodyne beatnote generated in the photodetector was frequency demodulated with an RF spectrum analyzer (Rohde&Schwarz FSV7).
Fig. 3 Left: schematic diagram of the experimental layout: EYDFA – erbium/ytterbium doped fiber amplifier, X0/Y0 – fiber couplers with appropriate split ratio, BPF – band pass filter for blocking pump radiation, AOM – acousto-optical modulator/frequency shifter, PD – photodiode. Mechanical chopper was used to periodically excite HCN molecules. Right: frequency of the beatnote signal indicates changes in the optical path length when pump light is tuned to the center of the HCN absorption line (black trace), while no photo-acoustic signal from HCN molecules is observed when wavelength is tuned away from the molecular transition (red trace).
Figure 3 also shows FM demodulated beatnote signals measured when pump wavelength is tuned on- and off-line respectively. When the pump laser is tuned to the center of HCN transition a periodic signal with its period matching the chopping frequency is clearly observed. This signal is not presented in the off-line measurement.
One can notice small residual periodic flat features visible in the off-line trace shown in Fig. 3. It has been identified that this residual signal is pump-wavelength-independent and it originates from heating/cooling of the optical components in the interferometric setup. As shown in Fig. 4 this spurious signal can be effectively eliminated when an intensity modulation of the pump laser is replaced with wavelength modulation (WM), which generates PA signal determined strictly by an absorption line shape. In order to implement WM detection the signal obtained after FM demodulation must be filtered with a narrow-band filer (e.g. lock-in amplifier) at harmonics of the modulation frequency (in this experiment 2 × fm component was measured) [14]. This experimental arrangement is shown in Fig. 4, with sample spectrum recorded as the pump wavelength was scanned across a HCN transition. For comparison an acquisition of a direct absorption spectrum was also performed in this setup by measuring the probe laser light intensity after the sample (see the gray trace in Fig. 4).
Fig. 4 Left: schematic diagram of the proposed heterodyne-based detection in photoacoustic spectroscopy setup with wavelength modulation. Right: 2f PAS spectrum (red) recorded for fm = 500 Hz, P = 1 W (measured at the output of the EYDFA), single data point corresponds to acquisition time of approximately 50 ms. For comparison a tunable diode laser absorption spectroscopy (TDLAS) of the same transition is also shown (gray trace).
3.2 Open-path sensing
As mentioned earlier, optical-based detection of PA signal provides unique opportunities of implementation of different sensing configurations. In order to demonstrate these capabilities a setup with a pump source operating at ~2 µm was constructed. This spectral region contains relatively strong transitions of CO2 and enables application of thulium doped fiber amplifiers (TDFA) to boost the optical power of conventional diode lasers. An optical layout of the system is shown in Fig. 5. A DFB laser diode operating at ~1550 nm was used as a probe source. The pump and the probe laser beams were combined using a dichroic mirror (DM) and propagated together for a distance of approximately 25 cm (both beams were ~3 mm in diameter). A second dichroic mirror was used to reject the pump beam whereas the probe beam was coupled back into an optical fiber based interferometer setup. The setup was mounted onto a breadboard that was placed in a sealed box, which was subsequently filled with CO2/air mixture. The concentration of CO2 in the container was estimated at a level of 50% using TDLAS measurement and line-by-line fitting of the measured spectrum. Figure 6 demonstrates an impact of experimental parameters (pump laser power, and WM depth) on the recorded PA spectra. As the WM depth is increased the 2f spectra remain baseline-free and their shape changes in a similar way as commonly expected in wavelength modulation spectroscopy (WMS) systems [14]. The optical power of the pump source has direct impact on the signal amplitude (approximately linear dependence).
Fig. 5 Schematic diagram of PAS measurement using a 2-µm tunable fiber laser as the pump source, and a Mach-Zehnder fiber interferometer at 1.55 µm for path-length modulation signal retrieval. TDFA – thulium doped fiber amplifier, DM – dichroic mirror, AOM – acousto-optical modulator (Ω = 40 MHz), PD – photodiode.
Fig. 6 (left) Six consecutive 2f WM photoacoustic spectral scans registered for 50% CO2 at atmospheric pressure within a 25-cm pathlength at different values of the pump laser WM depth (left graph, modulation frequency fm = 250 Hz, “pump” optical power 330 mW); (right) Five consecutive 2f WM photoacoustic spectra recorded for different optical powers of the pump source (fm = 250 Hz, 6 mA modulation depth). Each measurement point was averaged for 500 ms.
3.3 Gas sensing with absorber localization
In the previous setup shown in Fig. 5 the pump and the probe beams are co-linear, which results in a path-integrated spectrum of the sample over the whole interaction distance. However this measurement technique can be also arranged so that the pump and probe beams are overlapped (cross each other) only at a certain location in space. In such a configuration only the PAS signal generated in a small volume where the two beams overlap will be measured. With appropriate spatial control of these two beams this configuration can be used for direct three-dimensional measurement of gas distribution without a need for any special tomographic reconstruction algorithms and signal processing [15–17].
Figure 7 shows a proof of concept optical layout of the setup to demonstrate spatial gas distribution measurement. The location where the pump and probe beams (both with ~3mm diameter) cross each other is set by translating the crossing pump beam along the axis of the probe beam (performed by translating its launching fiber collimator along an optical rail). The measured sample was pure CO2 in a glass cell (25 mm diameter) placed in a fixed location along the probe beam (the position of the cell was such that the beams were passing through its cylindrical side wall). The pump source wavelength was tuned to the center of the CO2 transition near 2003 nm and a fiber amplifier was used to obtain approximately 300 mW of optical power at the output of the launching collimator.
Fig. 7 PAS configuration for spatially resolved measurement. AOM – acousto optical modulator, PD – photodiode. Gas cell filled with pure CO2 at atmospheric pressure was placed in optical path of the 1.55-µm Mach-Zehnder interferometer. Clear PA signal was registered only when the 2 µm pump beam illuminated the gas cell (right graph).
A measured signal amplitude recorded as a function of the pump beam displacement is shown in Fig. 7. When the interaction point of the two beams falls outside the gas cell, the measured PA signal amplitude is around zero. However, when the probed volume is located inside the cell, PA signal is clearly detected. The position and orientation of the collimated beams defines the location of the probing volume, and thus with a relatively simple beam steering configuration this location can be translated in space to accomplish three-dimensional measurement of spatial gas distribution.
3.3 Detection limit
A gas cell (20-cm long) with pure CO2 at 740 Torr was used to estimate the detection limit of the measurement technique. A laser diode at ~1573 nm was used as a pump source. Additional erbium-doped fiber amplifier boosted the output power to approximately 300 mW. With this wavelength relatively weak absorption lines located in the R-branch of the 2v1 + 2v2 + v3 combination band of CO2 could be targeted. For the transition near 1573 nm peak absorption was below 2.5%, which assured a linear regime of the Beer-Lambert law. The laser wavelength was scanned through the CO2 line to record its signal amplitude. Subsequently, a gas cell was removed to perform long-term zero-gas measurement (with acquisition rate of 10 Hz) used for Allan deviation analysis. Figure 8 shows noise equivalent absorption (NEA) vs. integration time. For the integration time of one second NEA reaches 10−4 level which corresponds to normalized noise equivalent absorption coefficient (NNEA) of 1.3 × 10−6 cm−1W/Hz0.5 (which is at similar level as recently demonstrated in [13] for standard interferometric-based photothermal spectroscopy). This is achieved without using any acoustic resonator for signal enhancement, with relatively large beam size (diameter of 3 mm) and in configuration that allows increasing the optical path with a multi pass cell. Plot shows behavior similar to white noise even for long averaging times. This performance results from two properties of the presented method: baseline free nature and immunity to optical fringes (this phenomena is also known from other techniques which rely on gas sample modulation, e.g. from Faraday modulation spectroscopy [18]).
4. Conclusions
In this paper a new heterodyne-based method for optical measurement of photoacoustic signal was presented. The direct optical detection allows for sensing configurations that cannot be easily implemented using standard photoacoustic spectroscopy approaches. Comparing to previously reported examples of interferometric-based signal detection in photoacoustic or photothermal spectroscopy, presented heterodyne-based technique does not require the interferometer to be perfectly balanced and stabilized. This simplifies the optical design allowing for new sensing arrangements. In this work we have demonstrated photoacoustic spectroscopy of CO2 at 2003 nm performed in open path, without using any gas or acoustic cells. Unique capabilities of presented technique are also presented in novel gas sensing configuration that enables three-dimensional spatial gas distribution measurement.
This technique can be further explored using mid-infrared sources (quantum and interband cascade lasers or differential frequency generation-based) to target the strongest fundamental molecular transitions in the 3-µm to 10-µm spectral region, still with convenient signal detection at shorter wavelengths (e.g. in visible or telecom range). Fundamentally, the presented approach allows for very flexible selection of the modulation frequency (it might be limited if acoustical resonator is applied for signal enhancement) and is strongly immune to optical fringes (which usually limit the performance of standard, absorption-based methods). Further development of this measurement technique will focus on extending optical paths using multi-pass cells or hollow-core photonic crystal fiber for further enhancement of detection sensitivities.
Funding
National Science Centre, Poland (DEC-2014/14/M/ST7/00866).
Acknowledgment
MN acknowledges a scholarship for young scientists from Polish Ministry of Science and Higher Education.
References and links
1. L. B. Kreuzer, “Ultralow Gas Concentration Infrared Absorption Spectroscopy,” J. Appl. Phys. 42(7), 2934–2943 (1971). [CrossRef]
2. A. A. Kosterev, Y. A. Bakhirkin, and F. K. Tittel, “Ultrasensitive gas detection by quartz-enhanced photoacoustic spectroscopy in the fundamental molecular absorption bands region,” Appl. Phys. B 80(1), 133–138 (2005). [CrossRef]
3. Y. 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] [PubMed]
4. B. A. Paldus, T. G. Spence, R. N. Zare, J. Oomens, F. J. M. Harren, D. H. Parker, C. Gmachl, F. Cappasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, and A. Y. Cho, “Photoacoustic spectroscopy using quantum-cascade lasers,” Opt. Lett. 24(3), 178–180 (1999). [CrossRef] [PubMed]
5. P. Patimisco, G. Scamarcio, F. K. Tittel, and V. Spagnolo, “Quartz-Enhanced Photoacoustic Spectroscopy: A Review,” Sensors (Basel) 14(4), 6165–6206 (2014). [CrossRef] [PubMed]
6. V. Spagnolo, P. Patimisco, S. Borri, G. Scamarcio, B. E. Bernacki, and J. Kriesel, “Mid-infrared fiber-coupled QCL-QEPAS sensor,” Appl. Phys. B 112(1), 25–33 (2013). [CrossRef]
7. M. A. Owens, C. C. Davis, and R. R. Dickerson, “A Photothermal Interferometer for Gas-Phase Ammonia Detection,” Anal. Chem. 71(7), 1391–1399 (1999). [CrossRef] [PubMed]
8. C. Davis, “Trace detection in gases using phase fluctuation optical heterodyne spectroscopy,” Appl. Phys. Lett. 36(7), 515–518 (1980). [CrossRef]
9. F. Yang, Y. Tan, W. Jin, Y. Lin, Y. Qi, and H. L. Ho, “Hollow-core fiber Fabry-Perot photothermal gas sensor,” Opt. Lett. 41(13), 3025–3028 (2016). [CrossRef] [PubMed]
10. W. Jin, Y. Cao, F. Yang, and H. L. Ho, “Ultra-sensitive all-fibre photothermal spectroscopy with large dynamic range,” Nat. Commun. 6, 6767 (2015). [CrossRef] [PubMed]
11. Y. Lin, W. Jin, F. Yang, J. Ma, C. Wang, H. L. Ho, and Y. Liu, “Pulsed photothermal interferometry for spectroscopic gas detection with hollow-core optical fibre,” Sci. Rep. 6(1), 39410 (2016). [CrossRef] [PubMed]
12. Y. Lin, W. Jin, F. Yang, and C. Wang, “Highly sensitive and stable all-fiber photothermal spectroscopic gas sensor,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (2016) (Optical Society of America, 2016), STu4H.3. [CrossRef]
13. Z. Li, Z. Wang, F. Yang, W. Jin, and W. Ren, “Mid-infrared fiber-optic photothermal interferometry,” Opt. Lett. 42(18), 3718–3721 (2017). [CrossRef] [PubMed]
14. S. Schilt and L. Thévenaz, “Wavelength modulation photoacoustic spectroscopy: Theoretical description and experimental results,” Infrared Phys. Technol. 48(2), 154–162 (2006). [CrossRef]
15. Y. Li, H. Xu, R. Xue, X. Wang, Y. Ren, L. Wang, and J. Wang, “Path Concentration Distribution of Toluene using Remote Sensing FTIR and One-Dimensional Reconstruction Method,” J Environ Sci Health A Tox Hazard Subst Environ Eng 40(1), 183–191 (2005). [CrossRef] [PubMed]
16. D. C. Wolfe Jr and R. L. Byer, “Model studies of laser absorption computed tomography for remote air pollution measurement,” Appl. Opt. 21(7), 1165–1178 (1982). [CrossRef] [PubMed]
17. R. L. Byer and L. A. Shepp, “Two-dimensional remote air-pollution monitoring via tomography,” Opt. Lett. 4(3), 75–77 (1979). [CrossRef] [PubMed]
18. Y. Wang, M. Nikodem, E. Zhang, F. Cikach, J. Barnes, S. Comhair, R. A. Dweik, C. Kao, and G. Wysocki, “Shot-noise Limited Faraday Rotation Spectroscopy for Detection of Nitric Oxide Isotopes in Breath, Urine, and Blood,” Sci. Rep. 5(1), 9096 (2015). [CrossRef] [PubMed]
