REVIEW OF SCIENTIFIC INSTRUMENTS 86, 023111 (2015)
Quartz-enhanced photoacoustic spectroscopy sensor for ethylene detection with a 3.32 µm distributed feedback laser diode T. Nguyen Ba,1,2 M. Triki,1,2 G. Desbrosses,3 and A. Vicet1,2,a)
1
Université de Montpellier, IES, UMR 5214, F-34000 Montpellier, France CNRS, IES, UMR 5214, F-34000 Montpellier, France 3 LSTM, UMR 113, IRD, CIRAD, UM2, SupAgro, Université Montpellier 2, Place E. Baaillon, F-34095 Montpellier, France 2
(Received 5 November 2014; accepted 10 February 2015; published online 25 February 2015) An antimonide distributed feedback quantum wells diode laser operating at 3.32 µm at near room temperature in the continuous wave regime has been used to perform ethylene detection based on quartz enhanced photoacoustic spectroscopy. An absorption line centered at 3007.52 cm−1 was investigated and a normalized noise equivalent absorption coefficient (1σ) of 3.09 10−7 cm−1 W Hz−1/2 was obtained. The linearity and the stability of the detection have been evaluated. Biological samples’ respiration has been measured to validate the feasibility of the detection setup in an agronomic environment, especially on ripening apples. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4913383] I. INTRODUCTION
Ethylene (C2H4) is an organic product produced in high quantities by the industry, through the process of cracking hydrocarbons from petroleum. Its production is increasing due to many applications, such as the preparation of plastics by polymerization, ethylene oxide production and ethylene glycol (antifreeze) fabrication by oxidation, and polystyrene fabrication by alkylation. Ethylene is also naturally produced by plants. It is a plant hormone involved in complex bio-regulation processes;1 it can change the flower gender of climateric fruits, such as apples, pears, melons, apricots, and bananas, as well as stimulate their maturation; the ripening of fruit can be adjusted by controlling the ethylene concentration in an air-controlled chamber. For example, during a few days of the maturation phase, the ethylene emission by a melon is approximately 100 nl g−1 h−1,2 which, related to a fruit total weight, is equal to approximately 100 ppmv per hour. A low concentration (few ppmv) is enough to activate the maturation processes. When a large quantity of fruits or vegetables is stored, it is necessary to regulate small ethylene amounts to stay below the ripening and degradation concentrations. As a consequence, there is an increasing demand for reliable and affordable ethylene sensor to control ethylene concentration.3 Gas chromatography is one of the most sensitive implemented methods for ethylene measurement:4 such detectors can measure concentrations of a few ppb in the atmosphere,5 but a qualification is needed from operating personnel, and consumables (chromatography columns, carrier gases, etc.) are mandatory.6 Electro-catalytic sensors are sometimes used but they show saturation or wear and degradation problems without enough selectivity or accuracy. Optical gas sensors can provide selectivity, accuracy, and sensitivity in ethylene sensing3 especially using lasers as excitation sources. Photoacoustic spectroscopy,7 using a high a)Electronic mail: [email protected]
0034-6748/2015/86(2)/023111/5/$30.00
power CO2 laser emitting at 10 µm, has been used to develop very efficient sensors dedicated to plants monitoring,3 with a very wide range of biological applications. Quartz-enhanced photoacoustic spectroscopy (QEPAS),8 an approach of photoacoustic detection, is a very good candidate for the fabrication of a small and mobile ethylene sensor, which could be used in the agri-food sector. It is based on the use of a Quartz Tuning Fork (QTF) as a sharply acoustic transducer, which converts an acoustic wave generated by gas absorption of the modulated laser source into an electrical signal due to its piezoelectric effect. The amplitude of the detected signal is proportional to the absorption coefficient, the quality factor of the QTF and the excitation power of the laser.9 Commercial QTFs usually have a resonant frequency of 32.76 kHz and a high Q-factor (∼100 000 under vacuum and ∼7000 at atmospheric pressure in ambient air), which make them highly frequency-selective. Moreover, the small QTF dimensions (∼2 mm3) allow the realization of compact gas cell. QEPAS has already been used to perform ethylene detection. A 1.62 µm distributed feedback (DFB) laser was used10 with an average optical power of 15 mW. A noise equivalent signal of 4 ppmv was achieved for 0.7 s data acquisition with a 2f-detection on an absorption peak at 6177.14 cm−1, far from H2O and CO2 interference. In this paper, we report our results on a prototype bench devoted to C2H4 detection based on the QEPAS technique using an antimonide laser diode emitting at 3.32 µm. This wavelength range in the ν11 band can address 5 to 8 times stronger absorption lines than at 1.62 µm, while the lasers properties, especially power and spectral purity, made strong progresses these last five years.11
II. LASER CHARACTERIZATION
An antimonide DFB diode laser emitting at 3325 nm was used in this setup. It was grown by molecular beam epitaxy on a GaSb substrate and is composed of four InGaAsSb quantum
86, 023111-1
© 2015 AIP Publishing LLC
023111-2
Nguyen Ba et al.
FIG. 1. Light-intensity-voltage curves of the DFB laser, T = 4 ◦C. Inset: T0 determination from 3 to 8 ◦C.
wells embedded in AlGaInAsSb barriers following the structure described in Ref. 11. The DFB devices were processed from this sample by Nanoplus GMBH. A complex DFB coupling is obtained by the deposition of a first order metallic grating on both sides of the laser ridge. A similar device has already been used on QEPAS setups to perform CH4 detection.12,13 Figure 1 shows the main properties of the component. The operating temperature was 4 ◦C in continuous wave regime. It is temperature-controlled with a Peltier cooler and mounted on a heat dissipater. The temperature is precisely controlled without drift, even working in ambient temperature. At 4 ◦C, the threshold current was 67 mA, the built-in potential 1 V, and the serial resistance 4.5 Ω. Within the 3 to 8 ◦C range, the value of the characteristic temperature T0 was evaluated to be 37 K. While this is a low value, it must be taken into consideration that it was estimated within a very small temperature range. The maximum output power reached 3 mW per facet for an injected current of 140 mA. Figure 2 gives the emitted wavelength versus temperature (Fig. 2(a)) and current (Fig. 2(b)), recorded with a grating spectrometer of 0.5 nm spectral resolution (Cornerstone 74 125). From 276 to 282 K, at 120 mA the laser can be continuously tuned from 3.3238 to 3.3254 µm, which gives a temperature tuning of 0.26 nm/K, typical for this type of
Rev. Sci. Instrum. 86, 023111 (2015)
component. The current tuning was 0.025 nm/mA, which can be mainly explained by thermal effects under current injection leading to the shift of the cavity laser modes in relation to the refractive index variation. The side mode suppression ratio was approximately 30 dB within the observed tuning range. To complete the characterization of the component, the laser light absorption was recorded on an InGaAs photodiode (PD) through a reference cell (Wavelength References, Inc., L = 2 cm, P = 740 Torr) filled with pure ethylene. Several absorption lines were identified using the HITRAN database,14 and we decided to focus on a line centered at 3007.52 cm−1, with a line strength S = 8.94 × 10−21 cm−1/mol cm−2. Figure 3 shows a simulation (HITRAN PC) of the 1 cm-transmission of a mixture composed of 2.5% of H2O, 380 ppmv of CO2, and 100 ppmv of C2H4. As the working spectral range is mainly free of CO2 and H2O lines, only a small water line appears on the plot, but out of the laser tuning range. The other lines are all C2H4. An arrow shows the studied line.
III. QEPAS SETUP
A typical QEPAS scheme was proposed on this setup (Fig. 4). It consists of a 3.325 µm CW DFB laser, a (L1) black diamond antireflection coated (3-5 µm) aspheric lens (f = 4 mm), and an acoustic detection module (ADM) inside which contains a QTF, one or two acoustic micro-resonators (mRs) and a transimpedance amplifier (TA). The DFB diode laser is mounted on a xyz micro positioning sub-mount. The device is driven by a current source (Profile, model ITC-502) and its temperature is regulated by a proportional integral differential (PID) temperature controller (ILX model LDT-5901B). In order to realize the detection of ethylene absorption line that was previously identified (σ0 = 3007.52 cm−1), the laser temperature has to be stabilized at 4 ◦C and its injected current is tuned from 80 to 120 mA. Under these conditions, the laser power is ∼1.5 mW. However, the optical power transmitted to the QTF is only 0.46 mW because a large part of the laser power is lost travelling from the component to the QTF. First, only 64% of the emitted light is collected by the focalization black diamond lens (L1) because of the high divergence of the laser beam (vertical half width at half maximum >50◦). Then, 35% of the collected light is absorbed by L1. Finally, 26% of the remaining light
FIG. 2. Tuning properties of the DFB laser. (a): Emitted wavelength vs laser temperature for a given current I = 120 mA. (b): Emitted wavelength vs injected current for a given temperature T = 4 ◦C.
023111-3
Nguyen Ba et al.
FIG. 3. HITRAN-PC simulation of a mixture composed of 2.5% H2O, 380 ppmv CO2, and 100 ppmv C2H4. L = 1 cm, P = P atm. Only one water line can be seen here. The studied ethylene line is also shown. The laser tuning range is shown for a temperature of 4 ± 1 ◦C.
is absorbed by the sapphire window of the QEPAS cell. As a consequence, only 31% of the total emitted light (∼0.46 mW) is transmitted to the QTF. Usually, in order to improve the detection sensitivity of the QEPAS sensor, mRs, made of glass or stainless steel, are used in “on-beam” or “off-beam” configuration.8,9,15 In the “onbeam” configuration, two mRs of the same length are used, placed on each side of the QTF. In the “off-beam” configuration, a mR with a small slit in the middle can be used and placed beside the QTF. While the “off-beam” configuration gives an easier optical alignment, the “on-beam” configuration gives a higher enhancement factor on the QEPAS signal. We have chosen to work with on-beam configuration. The two stainless mRs (4.4 mm-long with an inner diameter of
Rev. Sci. Instrum. 86, 023111 (2015)
0.5 mm) are mounted on both sides of the QTF, very close to the external surface of the prongs: this distance does not exceed 50 µm (Fig. 4). The mRs are placed at the optimum position:16 0.7 mm from the top of the QTF prongs. At atmospheric pressure, the resonant frequency of the QTF with 2 mR was measured at 32.751 kHz with a quality factor Q ∼ 6000 following the typical experimental setup described in Ref. 17. Using these mRs gives an increase of a factor of 34 for 5000 ppmv concentration measurement without and with them. The acoustic detection module is a compact gas cell equipped with gas inlet and outlet, and two sapphire windows. L1 is used to focus the laser beam between the two prongs of the QTF. The distance from the laser to the plane of the lens is 4.15 mm, while the lens-QTF distance is 110 mm. The total beam waist between the two prongs of the QTF was evaluated to be 130 µm. The QTF dimensions are 6 mm × 1.5 mm, with a thickness of 0.3 mm; each prong is 3.8 mm long and 0.6 mm wide. For sensitive C2H4 measurements, wavelength modulation spectroscopy (WMS) with 2nd harmonic detection (2fdetection) was used with a lock-in amplifier (EG&G Instruments 7260). The laser was first fed with an initial current of 80 mA, then a 40 mA scan (0.1 Hz) was added to tune the emitted wavelength through the targeted absorption line. Wavelength modulation at half the QTF resonance frequency (fm = f0/2 ∼ 16.3 kHz) was implemented by applying a sinusoidal dither to the direct current ramp of the DFB laser. A current modulation amplitude of ∆I = 16 mA was applied on the injected current through an electronic filter used to convert the voltage modulation from the lock-in internal oscillator (∆V = 3 V) to an intensity modulation for the laser. The time constant of the lock-in amplifier was fixed at 1 s. In photo acoustic spectroscopy, when the gas molecules absorb the optical energy deposited by the laser, an acoustic wave is generated due to thermal relaxation processes. The interaction of this acoustic wave with the QTF prongs excites the fundamental piezoelectric mode, which results in a
FIG. 4. Diagram of the experimental setup. (DL) Diode Laser, (L) lens, (QTF) Quartz Tuning Fork, (mR) microresonator, (T.A) Transimpedance Amplifier, (M) Mirror, (Ref cell) reference cell, (BS) Beam Splitter, (PD) photodiode, (PC) Personal Computer.
023111-4
Nguyen Ba et al.
weak current in the device.18 The current is then converted into voltage and amplified by a homemade transimpedance amplifier with a 10 MΩ feedback resistor. The amplified signal at fm is collected by means of the lock-in amplifier and the data are recorded via an oscilloscope to a computer. Another important part of the optical setup is the direct absorption path. After passing through the ADM, the DFB laser beam is focused on an InAs PD placed after the 2 cm-long C2H4 reference cell. The photodiode is connected to a transimpedance amplifier and an oscilloscope. This part enables us to spectrally localize the best absorption line of the targeted gas and allows a first alignment of the QEPAS system. To realize a fine optical alignment, a pellicle beam splitter (B. S.) and a flip gold mirror (M) are used to let the visible beam emitted by a red laser diode follow the same optical path as the infrared beam of the 3.3 µm DFB laser. A visualization system is used to observe both the emitting facet of the infrared laser and the red laser spot on this facet after passing through the QTF prongs.
IV. RESULTS
Figure 5 gives several 2 f -signals recorded on the identified line. The time constant of the lock-in amplifier was set to 1 s (18 dB/octave) and the signals were averaged 10 times. Measurements were taken using calibrated C2H4/N2 gas mixtures obtained with a gas dilutor (GasMix AIOLOS from AlyTech). All measurements performed in this work were realized at atmospheric pressure. The gas flow is controlled during the calibration procedure with the dilutor, but the measurements are performed at atmospheric pressure without specific stabilization. The calibration curve (Fig. 6) shows a very good linearity of the signal with the concentrations. The uncertainty has been estimated with the signal baseline amplitude giving the noise level (about 80 ppmv). For those 10 s measurements, the normalized noise equivalent absorption (NNEA) can be estimated to 3.09 × 10−7 cm−1 W Hz−1/2. To evaluate the signal stability and the best achievable sensitivity of the system, we have recorded the Allan plot19 (Fig. 7). An acquisition was made each 1.5 s, with a 1 Hz current ramp and 100 ms time constant (slope 18 dB/octave)
FIG. 5. 2 f QEPAS signal given by calibrated gas mixtures composed of pure nitrogen and different C2H4 concentrations.
Rev. Sci. Instrum. 86, 023111 (2015)
FIG. 6. Calibration curve: 2 f signal recorded versus C2H4/N2 calibrated concentrations. The uncertainty is estimated with the signal baseline, corresponding to a signal to noise ratio of 1. The noise level corresponding to the baseline fluctuations is approximately 80 ppmv.
on the lock-in amplifier, without signal averaging. Under these conditions, a detection limit of 63 ppmv could be achieved for a 25 s averaging time, corresponding to a NNEA = 3.89 × 10−7 cm−1 W Hz−1/2 which is very consistent with the previously given value. Regarding measurements with 10 s averaging time, the relevant concentration limit is 80 ppmv, which is perfectly consistent with previous measurements. Measurements at 1.62 µm10 gave a detection limit of 4 ppmv in 0.7 s. The absorption line used in this previous study was not exhaustively descripted; its line strength was not given. However, taking into account the potentially stronger line used at 3.32 µm, we were supposed to reach a lower detection limit, which here is not the case. This fact can mainly be explained by the lower power of the laser source used here: 0.46 mW of the emitted power is deposited between the prongs of the fork while the 1.62 µm laser had an emitted power of 15 mW. To evaluate the sensor feasibility on a real biological sample, we have measured the ethylene production by an apple. The fruit was enclosed in a glass previously filled with pure
FIG. 7. Allan plot given by a mixture of 1% C2H4 and 99% N2. The 40 mA current ramp is applied in 1 s; the lock-in time constant is 100 ms.
023111-5
Nguyen Ba et al.
Rev. Sci. Instrum. 86, 023111 (2015)
power, one may improve the sensitivity of the detection using more powerful devices on stronger lines. Then, to reach stronger photoacoustic signal and decrease the detection limit of our setup, we plan to use a new type of buried DFB diode lasers20 that will be transposed on 3.3 µm devices. This work is also in progress, in order to develop a more compact system, to meet the needs in the agri-food sector. ACKNOWLEDGMENTS
FIG. 8. 2 f -QEPAS signal given by an apple.
This work is supported by the French National Research Agency within the NexCILAS international ANR/NSF Project No. (2011 NS09-002-01) and within the MIDAS ANR Project No. (2011 NANO-028-01), and the NUMEV Labex (French Laboratory of Excellence). 1F.
◦
nitrogen during one night in a fridge (5 C) to slow down the ethylene emission process, and then it was maintained 4 h out of the fridge at room temperature before starting the measurements. A gas volume corresponding to the ADM volume was extracted using a syringe from the gas surrounding the fruit. The gas mixture was then composed of nitrogen plus the exhausted species from the fruit. There was no ambient air in the mixture. The strong QEPAS signal given by the apple was compared to the calibration curve and to a calibrated QEPAS signal obtained with a calibrated C2H4/N2 gas mixture (Fig. 8). A very good fit is obtained with a 90 ppmv calibrated C2H4/N2 mixture, corresponding to the fruit emission. The shape of these signals is a bit different from those shown in Figure 5 because they concern concentrations reaching the detection limit of our setup, they are more sensitive to distortions’ dues to noise and background variation.
V. CONCLUSION
We have reported on the realization of an ethylene sensor based on quartz enhanced photoacoustic spectroscopy operating with a 3.32 µm continuous wave DFB laser diode. The experimental results show that such a sensor could be exploited either in the ethylene production industry or in agrifood production. The small size of the sensor along with the simplicity of the technique is the main advantages of the detection system, which can be adapted to any laser source. Here, with only 0.46 mW focused between the two prongs of the tuning fork, we have obtained a NNEA value of 2.7 × 10−7 cm−1 W Hz−1/2. As the QEPAS technique, as well as every photo acoustic technique, is strongly dependent on laser
Vandenbussche, I. Vaseva, K. Vissenberg, and D. Van Der Straeten, New Phytol. 194, 895 (2012). 2J. C. Pech, M. Bouzayen, and A. Latché, Plant Sci. 175, 114 (2008). 3S. M. Cristescu, J. Mandon, D. Arslanov, J. De Pessemier, C. Hermans, and F. J. M. Harren, Ann. Bot. 111, 347 (2013). 4J. R. Gorny and A. A. Kader, Perishables Handl. 97, 25 (1999). 5P. Pham-Tuan, J. Vercammen, and C. Davos, J. Chromatogr. A 868, 249 (2000). 6E. Wahl, S. Tan, S. Koulikov, B. Kharlamov, C. Rella, E. Crosson, D. Biswell, and B. Paldus, Opt. Express 14, 1673 (2006). 7F. J. M. Harren, J. Mandon, and S. M. Cristescu, Encyclopedia of Analytical Chemistry (John Wiley & Sons, Hoboken, NJ, 2012). 8A. A. Kosterev, Y. A. Bakhirkin, R. F. Curl, and F. K. Tittel, Opt. Lett. 27, 1902 (2002). 9R. Lewicki, G. Wysocki, A. A. Kosterev, and F. K. Tittel, Appl. Phys. B 87, 157 (2006). 10S. Schilt, A. A. Kosterev, and F. K. Tittel, Appl. Phys. B 95, 813 (2008). 11L. Naehle, S. Belahsene, M. V. Edlinger, M. Fischer, G. Boissier, P. Grech, G. Narcy, A. Vicet, Y. Rouillard, J. Koeth, and L. Worschech, Electron. Lett. 47, 46 (2011). 12M. Jahjah, A. Vicet, and Y. Rouillard, Appl. Phys. B 106, 483 (2012). 13H. Yi, W. Chen, A. Vicet, Z. Cao, X. Gao, T. Nguyen-Ba, M. Jahjah, Y. Rouillard, L. Nahle, and M. Fischer, Appl. Phys. B 116, 423 (2014). 14L. 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. Le Roy, 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, J. Quant. Spectrosc. Radiat. Transfer 130, 4 (2013). 15H. Yi, W. Chen, X. Guo, S. Sun, K. Liu, T. Tan, W. Zhang, and X. Gao, Appl. Phys. B 108, 361 (2012). 16P. Patimisco, G. Scamarcio, F. K. Tittel, and V. Spagnolo, Sensors 14(4), 6165 (2014). 17A. A. Kosterev, F. K. Tittel, D. V. Serebryakov, A. L. Malinovsky, and I. V. Morozov, Rev. Sci. Instrum. 76, 043105 (2005). 18J. M. Friedt and E. Carry, Am. J. Phys. 75, 415 (2007). 19P. Werle, R. Mücke, and F. Slemr, Appl. Phys. B 57, 131 (1993). 20Q. Gaimard, L. Cerutti, R. Teissier, and A. Vicet, Appl. Phys. Lett. 104, 161111 (2014).
Review of Scientific Instruments is copyrighted by AIP Publishing LLC (AIP). Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. For more information, see http://publishing.aip.org/authors/rights-and-permissions.
Quartz-enhanced photoacoustic spectroscopy sensor for ethylene detection with a 3.32 µm distributed feedback laser diode T. Nguyen Ba,1,2 M. Triki,1,2 G. Desbrosses,3 and A. Vicet1,2,a)
1
Université de Montpellier, IES, UMR 5214, F-34000 Montpellier, France CNRS, IES, UMR 5214, F-34000 Montpellier, France 3 LSTM, UMR 113, IRD, CIRAD, UM2, SupAgro, Université Montpellier 2, Place E. Baaillon, F-34095 Montpellier, France 2
(Received 5 November 2014; accepted 10 February 2015; published online 25 February 2015) An antimonide distributed feedback quantum wells diode laser operating at 3.32 µm at near room temperature in the continuous wave regime has been used to perform ethylene detection based on quartz enhanced photoacoustic spectroscopy. An absorption line centered at 3007.52 cm−1 was investigated and a normalized noise equivalent absorption coefficient (1σ) of 3.09 10−7 cm−1 W Hz−1/2 was obtained. The linearity and the stability of the detection have been evaluated. Biological samples’ respiration has been measured to validate the feasibility of the detection setup in an agronomic environment, especially on ripening apples. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4913383] I. INTRODUCTION
Ethylene (C2H4) is an organic product produced in high quantities by the industry, through the process of cracking hydrocarbons from petroleum. Its production is increasing due to many applications, such as the preparation of plastics by polymerization, ethylene oxide production and ethylene glycol (antifreeze) fabrication by oxidation, and polystyrene fabrication by alkylation. Ethylene is also naturally produced by plants. It is a plant hormone involved in complex bio-regulation processes;1 it can change the flower gender of climateric fruits, such as apples, pears, melons, apricots, and bananas, as well as stimulate their maturation; the ripening of fruit can be adjusted by controlling the ethylene concentration in an air-controlled chamber. For example, during a few days of the maturation phase, the ethylene emission by a melon is approximately 100 nl g−1 h−1,2 which, related to a fruit total weight, is equal to approximately 100 ppmv per hour. A low concentration (few ppmv) is enough to activate the maturation processes. When a large quantity of fruits or vegetables is stored, it is necessary to regulate small ethylene amounts to stay below the ripening and degradation concentrations. As a consequence, there is an increasing demand for reliable and affordable ethylene sensor to control ethylene concentration.3 Gas chromatography is one of the most sensitive implemented methods for ethylene measurement:4 such detectors can measure concentrations of a few ppb in the atmosphere,5 but a qualification is needed from operating personnel, and consumables (chromatography columns, carrier gases, etc.) are mandatory.6 Electro-catalytic sensors are sometimes used but they show saturation or wear and degradation problems without enough selectivity or accuracy. Optical gas sensors can provide selectivity, accuracy, and sensitivity in ethylene sensing3 especially using lasers as excitation sources. Photoacoustic spectroscopy,7 using a high a)Electronic mail: [email protected]
0034-6748/2015/86(2)/023111/5/$30.00
power CO2 laser emitting at 10 µm, has been used to develop very efficient sensors dedicated to plants monitoring,3 with a very wide range of biological applications. Quartz-enhanced photoacoustic spectroscopy (QEPAS),8 an approach of photoacoustic detection, is a very good candidate for the fabrication of a small and mobile ethylene sensor, which could be used in the agri-food sector. It is based on the use of a Quartz Tuning Fork (QTF) as a sharply acoustic transducer, which converts an acoustic wave generated by gas absorption of the modulated laser source into an electrical signal due to its piezoelectric effect. The amplitude of the detected signal is proportional to the absorption coefficient, the quality factor of the QTF and the excitation power of the laser.9 Commercial QTFs usually have a resonant frequency of 32.76 kHz and a high Q-factor (∼100 000 under vacuum and ∼7000 at atmospheric pressure in ambient air), which make them highly frequency-selective. Moreover, the small QTF dimensions (∼2 mm3) allow the realization of compact gas cell. QEPAS has already been used to perform ethylene detection. A 1.62 µm distributed feedback (DFB) laser was used10 with an average optical power of 15 mW. A noise equivalent signal of 4 ppmv was achieved for 0.7 s data acquisition with a 2f-detection on an absorption peak at 6177.14 cm−1, far from H2O and CO2 interference. In this paper, we report our results on a prototype bench devoted to C2H4 detection based on the QEPAS technique using an antimonide laser diode emitting at 3.32 µm. This wavelength range in the ν11 band can address 5 to 8 times stronger absorption lines than at 1.62 µm, while the lasers properties, especially power and spectral purity, made strong progresses these last five years.11
II. LASER CHARACTERIZATION
An antimonide DFB diode laser emitting at 3325 nm was used in this setup. It was grown by molecular beam epitaxy on a GaSb substrate and is composed of four InGaAsSb quantum
86, 023111-1
© 2015 AIP Publishing LLC
023111-2
Nguyen Ba et al.
FIG. 1. Light-intensity-voltage curves of the DFB laser, T = 4 ◦C. Inset: T0 determination from 3 to 8 ◦C.
wells embedded in AlGaInAsSb barriers following the structure described in Ref. 11. The DFB devices were processed from this sample by Nanoplus GMBH. A complex DFB coupling is obtained by the deposition of a first order metallic grating on both sides of the laser ridge. A similar device has already been used on QEPAS setups to perform CH4 detection.12,13 Figure 1 shows the main properties of the component. The operating temperature was 4 ◦C in continuous wave regime. It is temperature-controlled with a Peltier cooler and mounted on a heat dissipater. The temperature is precisely controlled without drift, even working in ambient temperature. At 4 ◦C, the threshold current was 67 mA, the built-in potential 1 V, and the serial resistance 4.5 Ω. Within the 3 to 8 ◦C range, the value of the characteristic temperature T0 was evaluated to be 37 K. While this is a low value, it must be taken into consideration that it was estimated within a very small temperature range. The maximum output power reached 3 mW per facet for an injected current of 140 mA. Figure 2 gives the emitted wavelength versus temperature (Fig. 2(a)) and current (Fig. 2(b)), recorded with a grating spectrometer of 0.5 nm spectral resolution (Cornerstone 74 125). From 276 to 282 K, at 120 mA the laser can be continuously tuned from 3.3238 to 3.3254 µm, which gives a temperature tuning of 0.26 nm/K, typical for this type of
Rev. Sci. Instrum. 86, 023111 (2015)
component. The current tuning was 0.025 nm/mA, which can be mainly explained by thermal effects under current injection leading to the shift of the cavity laser modes in relation to the refractive index variation. The side mode suppression ratio was approximately 30 dB within the observed tuning range. To complete the characterization of the component, the laser light absorption was recorded on an InGaAs photodiode (PD) through a reference cell (Wavelength References, Inc., L = 2 cm, P = 740 Torr) filled with pure ethylene. Several absorption lines were identified using the HITRAN database,14 and we decided to focus on a line centered at 3007.52 cm−1, with a line strength S = 8.94 × 10−21 cm−1/mol cm−2. Figure 3 shows a simulation (HITRAN PC) of the 1 cm-transmission of a mixture composed of 2.5% of H2O, 380 ppmv of CO2, and 100 ppmv of C2H4. As the working spectral range is mainly free of CO2 and H2O lines, only a small water line appears on the plot, but out of the laser tuning range. The other lines are all C2H4. An arrow shows the studied line.
III. QEPAS SETUP
A typical QEPAS scheme was proposed on this setup (Fig. 4). It consists of a 3.325 µm CW DFB laser, a (L1) black diamond antireflection coated (3-5 µm) aspheric lens (f = 4 mm), and an acoustic detection module (ADM) inside which contains a QTF, one or two acoustic micro-resonators (mRs) and a transimpedance amplifier (TA). The DFB diode laser is mounted on a xyz micro positioning sub-mount. The device is driven by a current source (Profile, model ITC-502) and its temperature is regulated by a proportional integral differential (PID) temperature controller (ILX model LDT-5901B). In order to realize the detection of ethylene absorption line that was previously identified (σ0 = 3007.52 cm−1), the laser temperature has to be stabilized at 4 ◦C and its injected current is tuned from 80 to 120 mA. Under these conditions, the laser power is ∼1.5 mW. However, the optical power transmitted to the QTF is only 0.46 mW because a large part of the laser power is lost travelling from the component to the QTF. First, only 64% of the emitted light is collected by the focalization black diamond lens (L1) because of the high divergence of the laser beam (vertical half width at half maximum >50◦). Then, 35% of the collected light is absorbed by L1. Finally, 26% of the remaining light
FIG. 2. Tuning properties of the DFB laser. (a): Emitted wavelength vs laser temperature for a given current I = 120 mA. (b): Emitted wavelength vs injected current for a given temperature T = 4 ◦C.
023111-3
Nguyen Ba et al.
FIG. 3. HITRAN-PC simulation of a mixture composed of 2.5% H2O, 380 ppmv CO2, and 100 ppmv C2H4. L = 1 cm, P = P atm. Only one water line can be seen here. The studied ethylene line is also shown. The laser tuning range is shown for a temperature of 4 ± 1 ◦C.
is absorbed by the sapphire window of the QEPAS cell. As a consequence, only 31% of the total emitted light (∼0.46 mW) is transmitted to the QTF. Usually, in order to improve the detection sensitivity of the QEPAS sensor, mRs, made of glass or stainless steel, are used in “on-beam” or “off-beam” configuration.8,9,15 In the “onbeam” configuration, two mRs of the same length are used, placed on each side of the QTF. In the “off-beam” configuration, a mR with a small slit in the middle can be used and placed beside the QTF. While the “off-beam” configuration gives an easier optical alignment, the “on-beam” configuration gives a higher enhancement factor on the QEPAS signal. We have chosen to work with on-beam configuration. The two stainless mRs (4.4 mm-long with an inner diameter of
Rev. Sci. Instrum. 86, 023111 (2015)
0.5 mm) are mounted on both sides of the QTF, very close to the external surface of the prongs: this distance does not exceed 50 µm (Fig. 4). The mRs are placed at the optimum position:16 0.7 mm from the top of the QTF prongs. At atmospheric pressure, the resonant frequency of the QTF with 2 mR was measured at 32.751 kHz with a quality factor Q ∼ 6000 following the typical experimental setup described in Ref. 17. Using these mRs gives an increase of a factor of 34 for 5000 ppmv concentration measurement without and with them. The acoustic detection module is a compact gas cell equipped with gas inlet and outlet, and two sapphire windows. L1 is used to focus the laser beam between the two prongs of the QTF. The distance from the laser to the plane of the lens is 4.15 mm, while the lens-QTF distance is 110 mm. The total beam waist between the two prongs of the QTF was evaluated to be 130 µm. The QTF dimensions are 6 mm × 1.5 mm, with a thickness of 0.3 mm; each prong is 3.8 mm long and 0.6 mm wide. For sensitive C2H4 measurements, wavelength modulation spectroscopy (WMS) with 2nd harmonic detection (2fdetection) was used with a lock-in amplifier (EG&G Instruments 7260). The laser was first fed with an initial current of 80 mA, then a 40 mA scan (0.1 Hz) was added to tune the emitted wavelength through the targeted absorption line. Wavelength modulation at half the QTF resonance frequency (fm = f0/2 ∼ 16.3 kHz) was implemented by applying a sinusoidal dither to the direct current ramp of the DFB laser. A current modulation amplitude of ∆I = 16 mA was applied on the injected current through an electronic filter used to convert the voltage modulation from the lock-in internal oscillator (∆V = 3 V) to an intensity modulation for the laser. The time constant of the lock-in amplifier was fixed at 1 s. In photo acoustic spectroscopy, when the gas molecules absorb the optical energy deposited by the laser, an acoustic wave is generated due to thermal relaxation processes. The interaction of this acoustic wave with the QTF prongs excites the fundamental piezoelectric mode, which results in a
FIG. 4. Diagram of the experimental setup. (DL) Diode Laser, (L) lens, (QTF) Quartz Tuning Fork, (mR) microresonator, (T.A) Transimpedance Amplifier, (M) Mirror, (Ref cell) reference cell, (BS) Beam Splitter, (PD) photodiode, (PC) Personal Computer.
023111-4
Nguyen Ba et al.
weak current in the device.18 The current is then converted into voltage and amplified by a homemade transimpedance amplifier with a 10 MΩ feedback resistor. The amplified signal at fm is collected by means of the lock-in amplifier and the data are recorded via an oscilloscope to a computer. Another important part of the optical setup is the direct absorption path. After passing through the ADM, the DFB laser beam is focused on an InAs PD placed after the 2 cm-long C2H4 reference cell. The photodiode is connected to a transimpedance amplifier and an oscilloscope. This part enables us to spectrally localize the best absorption line of the targeted gas and allows a first alignment of the QEPAS system. To realize a fine optical alignment, a pellicle beam splitter (B. S.) and a flip gold mirror (M) are used to let the visible beam emitted by a red laser diode follow the same optical path as the infrared beam of the 3.3 µm DFB laser. A visualization system is used to observe both the emitting facet of the infrared laser and the red laser spot on this facet after passing through the QTF prongs.
IV. RESULTS
Figure 5 gives several 2 f -signals recorded on the identified line. The time constant of the lock-in amplifier was set to 1 s (18 dB/octave) and the signals were averaged 10 times. Measurements were taken using calibrated C2H4/N2 gas mixtures obtained with a gas dilutor (GasMix AIOLOS from AlyTech). All measurements performed in this work were realized at atmospheric pressure. The gas flow is controlled during the calibration procedure with the dilutor, but the measurements are performed at atmospheric pressure without specific stabilization. The calibration curve (Fig. 6) shows a very good linearity of the signal with the concentrations. The uncertainty has been estimated with the signal baseline amplitude giving the noise level (about 80 ppmv). For those 10 s measurements, the normalized noise equivalent absorption (NNEA) can be estimated to 3.09 × 10−7 cm−1 W Hz−1/2. To evaluate the signal stability and the best achievable sensitivity of the system, we have recorded the Allan plot19 (Fig. 7). An acquisition was made each 1.5 s, with a 1 Hz current ramp and 100 ms time constant (slope 18 dB/octave)
FIG. 5. 2 f QEPAS signal given by calibrated gas mixtures composed of pure nitrogen and different C2H4 concentrations.
Rev. Sci. Instrum. 86, 023111 (2015)
FIG. 6. Calibration curve: 2 f signal recorded versus C2H4/N2 calibrated concentrations. The uncertainty is estimated with the signal baseline, corresponding to a signal to noise ratio of 1. The noise level corresponding to the baseline fluctuations is approximately 80 ppmv.
on the lock-in amplifier, without signal averaging. Under these conditions, a detection limit of 63 ppmv could be achieved for a 25 s averaging time, corresponding to a NNEA = 3.89 × 10−7 cm−1 W Hz−1/2 which is very consistent with the previously given value. Regarding measurements with 10 s averaging time, the relevant concentration limit is 80 ppmv, which is perfectly consistent with previous measurements. Measurements at 1.62 µm10 gave a detection limit of 4 ppmv in 0.7 s. The absorption line used in this previous study was not exhaustively descripted; its line strength was not given. However, taking into account the potentially stronger line used at 3.32 µm, we were supposed to reach a lower detection limit, which here is not the case. This fact can mainly be explained by the lower power of the laser source used here: 0.46 mW of the emitted power is deposited between the prongs of the fork while the 1.62 µm laser had an emitted power of 15 mW. To evaluate the sensor feasibility on a real biological sample, we have measured the ethylene production by an apple. The fruit was enclosed in a glass previously filled with pure
FIG. 7. Allan plot given by a mixture of 1% C2H4 and 99% N2. The 40 mA current ramp is applied in 1 s; the lock-in time constant is 100 ms.
023111-5
Nguyen Ba et al.
Rev. Sci. Instrum. 86, 023111 (2015)
power, one may improve the sensitivity of the detection using more powerful devices on stronger lines. Then, to reach stronger photoacoustic signal and decrease the detection limit of our setup, we plan to use a new type of buried DFB diode lasers20 that will be transposed on 3.3 µm devices. This work is also in progress, in order to develop a more compact system, to meet the needs in the agri-food sector. ACKNOWLEDGMENTS
FIG. 8. 2 f -QEPAS signal given by an apple.
This work is supported by the French National Research Agency within the NexCILAS international ANR/NSF Project No. (2011 NS09-002-01) and within the MIDAS ANR Project No. (2011 NANO-028-01), and the NUMEV Labex (French Laboratory of Excellence). 1F.
◦
nitrogen during one night in a fridge (5 C) to slow down the ethylene emission process, and then it was maintained 4 h out of the fridge at room temperature before starting the measurements. A gas volume corresponding to the ADM volume was extracted using a syringe from the gas surrounding the fruit. The gas mixture was then composed of nitrogen plus the exhausted species from the fruit. There was no ambient air in the mixture. The strong QEPAS signal given by the apple was compared to the calibration curve and to a calibrated QEPAS signal obtained with a calibrated C2H4/N2 gas mixture (Fig. 8). A very good fit is obtained with a 90 ppmv calibrated C2H4/N2 mixture, corresponding to the fruit emission. The shape of these signals is a bit different from those shown in Figure 5 because they concern concentrations reaching the detection limit of our setup, they are more sensitive to distortions’ dues to noise and background variation.
V. CONCLUSION
We have reported on the realization of an ethylene sensor based on quartz enhanced photoacoustic spectroscopy operating with a 3.32 µm continuous wave DFB laser diode. The experimental results show that such a sensor could be exploited either in the ethylene production industry or in agrifood production. The small size of the sensor along with the simplicity of the technique is the main advantages of the detection system, which can be adapted to any laser source. Here, with only 0.46 mW focused between the two prongs of the tuning fork, we have obtained a NNEA value of 2.7 × 10−7 cm−1 W Hz−1/2. As the QEPAS technique, as well as every photo acoustic technique, is strongly dependent on laser
Vandenbussche, I. Vaseva, K. Vissenberg, and D. Van Der Straeten, New Phytol. 194, 895 (2012). 2J. C. Pech, M. Bouzayen, and A. Latché, Plant Sci. 175, 114 (2008). 3S. M. Cristescu, J. Mandon, D. Arslanov, J. De Pessemier, C. Hermans, and F. J. M. Harren, Ann. Bot. 111, 347 (2013). 4J. R. Gorny and A. A. Kader, Perishables Handl. 97, 25 (1999). 5P. Pham-Tuan, J. Vercammen, and C. Davos, J. Chromatogr. A 868, 249 (2000). 6E. Wahl, S. Tan, S. Koulikov, B. Kharlamov, C. Rella, E. Crosson, D. Biswell, and B. Paldus, Opt. Express 14, 1673 (2006). 7F. J. M. Harren, J. Mandon, and S. M. Cristescu, Encyclopedia of Analytical Chemistry (John Wiley & Sons, Hoboken, NJ, 2012). 8A. A. Kosterev, Y. A. Bakhirkin, R. F. Curl, and F. K. Tittel, Opt. Lett. 27, 1902 (2002). 9R. Lewicki, G. Wysocki, A. A. Kosterev, and F. K. Tittel, Appl. Phys. B 87, 157 (2006). 10S. Schilt, A. A. Kosterev, and F. K. Tittel, Appl. Phys. B 95, 813 (2008). 11L. Naehle, S. Belahsene, M. V. Edlinger, M. Fischer, G. Boissier, P. Grech, G. Narcy, A. Vicet, Y. Rouillard, J. Koeth, and L. Worschech, Electron. Lett. 47, 46 (2011). 12M. Jahjah, A. Vicet, and Y. Rouillard, Appl. Phys. B 106, 483 (2012). 13H. Yi, W. Chen, A. Vicet, Z. Cao, X. Gao, T. Nguyen-Ba, M. Jahjah, Y. Rouillard, L. Nahle, and M. Fischer, Appl. Phys. B 116, 423 (2014). 14L. 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. Le Roy, 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, J. Quant. Spectrosc. Radiat. Transfer 130, 4 (2013). 15H. Yi, W. Chen, X. Guo, S. Sun, K. Liu, T. Tan, W. Zhang, and X. Gao, Appl. Phys. B 108, 361 (2012). 16P. Patimisco, G. Scamarcio, F. K. Tittel, and V. Spagnolo, Sensors 14(4), 6165 (2014). 17A. A. Kosterev, F. K. Tittel, D. V. Serebryakov, A. L. Malinovsky, and I. V. Morozov, Rev. Sci. Instrum. 76, 043105 (2005). 18J. M. Friedt and E. Carry, Am. J. Phys. 75, 415 (2007). 19P. Werle, R. Mücke, and F. Slemr, Appl. Phys. B 57, 131 (1993). 20Q. Gaimard, L. Cerutti, R. Teissier, and A. Vicet, Appl. Phys. Lett. 104, 161111 (2014).
Review of Scientific Instruments is copyrighted by AIP Publishing LLC (AIP). Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. For more information, see http://publishing.aip.org/authors/rights-and-permissions.
