Simultaneous Measurements of CO2 and CO Using a ...

Simultaneous Measurements of CO2 and CO Using a Single Distributed-Feedback (DFB) Diode Laser Near 1.57 lm at Elevated Temperatures TINGDONG CAI,* GUANGZHEN GAO, WEIDONG CHEN, GANG LIU, and XIAOMING GAO College of Physics and Electronic Engineering, Xuzhou Normal University, Xuzhou 221116, P.R. China (T.C.); Anhui Institute of Optics and Fine Mechanics, the Chinese Academy of Sciences, Hefei 230031, P.R. China (T.C., X.G.); Physical Department, Binzhou University, Binzhou 256600, P.R. China (G.G.); Laboratoire de Physicochimie de l’Atmosphe`re, Universite´ du Littoral Coˆte d’Opale 189A, Av. Maurice Schumann, 59140 Dunkerque, France (W.C.); and Basic Department of XuZhou Air Force College, Xuzhou 221005, P.R. China (G.L.)

A sensor using a single distributed-feedback (DFB) diode laser at 1.57 lm for the simultaneous measurement of CO2 and CO concentration at elevated temperatures is developed. A proper line pair near 6361.250 and 6361.344 cm1 is chosen based on absorption strength, separation of the two lines, and isolation from interference of neighboring transitions of the major combustion gases. The concentrations of CO2 and CO are inferred from their wavelength modulation spectroscopy (WMS) 1f-normalized absorption-based WMS-2f signal peak heights. The CO2 and CO concentration measurements are within 3.3% and 5% of the expected values over the full temperature range. Index Headings: Diode lasers; Elevated temperature; Wavelength modulation spectroscopy; CO2; CO; Concentration.

INTRODUCTION In recent years optical gas sensors based on tunable diode laser (TDL) absorption spectroscopy have become more and more important for monitoring and control of industrial combustion processes for improved efficiency, pollutant reduction, and product quality. These sensors can provide a fast, sensitive, nonintrusive, and reliable method for in situ measurements of multiple flow-field parameters such as concentration, temperature, pressure, and velocity in various harsh environments.1,2 The diode laser usually covers parts of the ro-vibrational bands of the species of interest in the nearinfrared or mid-infrared spectral regions. Semiconductor diode lasers are attractive sources for practical applications due to their compactness, availability, robustness, compatibility with fiber-optic technology, and relative ease of use. Studies of carbon dioxide and carbon monoxide have always been of immense importance in combustion. CO2 and CO are attractive target gases for hydrocarbon-fueled systems and their concentrations can be interpreted to indicate combustion efficiency. Many TDL sensors used to detect CO2 or CO have been developed for the mid- and nearinfrared regions,3–9 especially the near-infrared region, as shown in Fig. 1, although the absorbance of the second overtone and combination band in this region is relatively weak compared with the fundamental bands in the midinfrared region. The advances in detection methods (e.g., wavelength modulation spectroscopy, frequency modulation spectroscopy) over the past decade have greatly increased the sensitivity of diode laser absorption spectroscopy measureReceived 21 July 2010; accepted 5 October 2010. * Author to whom correspondence should be sent. E-mail: caitingdong@ 126.com. DOI: 10.1366/10-06074

108

Volume 65, Number 1, 2011

ments, such that detection of fractional absorbance on the order of 106 or better can be achieved.11,12 In addition, robust, compact, fiber-coupled, milliwatt, single-mode, telecommunications-grade diode lasers in the near-infrared region are commercially available. At present, the typical rapidtuning range of commercially available telecommunicationsquality single-mode distributed-feedback (DFB) diode lasers is 0 to 2 cm1. Hence, few of the sensors enable simultaneous measurements of multiple species and flow parameters along one line of sight by using a single DFB diode laser.13,14 Vertical-cavity surface-emitting lasers (VCELs) can be tuned continuously over extremely broad spectral intervals (. 30 cm1) because of their electrical heating effect associated with the large series resistance of distributed Bragg reflector mirrors and small cavity volume. However, when referring to gas detection at low pressures, this large current-tuning range of optical frequency was considered as a negative factor because of the corresponding large laser phase noise compared to the small transition line width at low pressures.15,16 Hence, VCELs are more suitable for gas detection in high-pressure environments. In this paper a fiber-optic diode laser sensor based on the second-harmonic detection of wavelength modulation spectroscopy (WMS-2f) for the simultaneous detection of CO2 and CO at elevated temperatures is presented. By proper spectral selection, both species can be measured with a single DFB diode laser around 1.57 lm. Candidate transitions are selected based on absorption strength, separation of the two lines, and isolation from interference of neighboring transitions of the major combustion gases. The laser modulation depth is optimized to maximize the WMS-2f signals of both lines and to simplify signal interpretation. In order to remove the transmission variation of the laser due to beam steering, mechanical misalignments, soot, and window fouling, CO2 and CO concentrations are measured by their WMS-1f-normalized absorption-based WMS-2f signal peak heights.

THEORY The transmission coefficient s(m) of monochromatic radiation through a uniform medium of length L [cm] can be obtained by the Beer–Lambert relation: It sðmÞ ¼ ¼ exp½aðmÞ ð1Þ I0 m Here It and I0 are the transmitted and incident laser intensities, respectively, and a(m) represents the spectral absorbance. For

0003-7028/11/6501-0108$2.00/0 Ó 2011 Society for Applied Spectroscopy

APPLIED SPECTROSCOPY

FIG. 1. Absorption line strengths of CO2 and CO at 1000 K (from HITRAN10).

FIG. 2. Simulated 2f signal of the selected pair near 6361.250 and 6361.344 cm1 using HITRAN 2004 for P ¼ 1 atm, T ¼ 1000 K, L ¼ 10 cm, and 2% CO2 and 2% CO in air.

an optically thin sample [a(m) , 0.1]: sðmÞ ¼ exp½aðmÞ ’ 1 aðmÞ ¼ 1 PxL

X

Sj ðTÞuj

ð2Þ

j

where P [atm] is the total pressure, x is the mole fraction of the absorbing species, L is the path length, Sj [cm2 atm1] is the line strength of the transition, uj [cm] Ris the line-shape function, which is normalized such that uj ðmÞdm [ 1, the subscript j denotes the jth absorption feature, and T [K] is the gas temperature. In wavelength-scanned WMS, it is a periodic even function in xt and can be expanded in a Fourier cosine series: ‘ X

s½¯m þ a cosðxtÞ ¼

Hk ð¯m; aÞcosðkxtÞ

ð3Þ

k¼0

and Pxi L p

Z

p

X

p

j

Sj uj ð¯m þ a cos hÞcos khdh

ð5Þ

where Hk is directly proportional to species concentration xi and path length L when the line-shape functions do not vary for the range of conditions found in the applications.17 Note that in addition to the absorption parameters, Hk also depends on the modulation depth a. This effect can be mitigated by choosing a proper modulation index m, which is defined as m¼

a Dm=2

S2f ð¯mÞ ’

GI¯0 GI¯0 H2 ð¯mÞ ¼ 2 2 Z PSðTÞxi L p uð¯m þ a cos hÞcos 2hdh p p ð7Þ

and S1f ð¯mÞ ’

where m¯ is the center frequency of the laser and a is the modulation depth. The components Hk ð¯m; aÞ can be described as Z Pxi L p X H0 ð¯m; aÞ ¼ Sj uj ð¯m þ a cos hÞdh ð4Þ 2p p j

Hk ð¯m; aÞ ¼

second Fourier component H2, and H0, which is close to unity, is the dominant term of the WMS-1f signal. Thus, the magnitude of the absorption-based WMS-2f signal, S2f ð¯mÞ and S1f ð¯mÞ, which are measured by a lock-in amplifier can be reduced as

ð6Þ

where Dm is the full width at half-maximum (FWHM) of the absorption line shape.18 For an isolated transition, it is well known that H2 is maximized while H1 and H3 are zero at line center. Hence, near the line center, the dominant term of the WMS-2f signal is the

GI¯0 i0 2

ð8Þ

where G is the optical-electrical gain of the detection system, I¯0 is the average laser intensity at m¯ , and i0 is the linear intensity modulation amplitude (i0 is the intensity amplitude around the average laser intensity I¯0 ). The major contribution to the 1f signal is generated by the inherent linear intensity modulation i0 cos(xt þ w1) when the injection current is modulated.17,19 In this analysis, an appropriate assumption for the hardware and modulation parameters is employed, so the laser intensity and wavelength are assumed to vary linearly with injection current. Hence, the contributions from absorption to the 1f signal can be neglected and the magnitude of the 1f signal can be approximated by the 1f signal with no absorption as shown in Eq. 8. From Eqs. 7 and 8, the WMS-1f-normalized absorptionbased WMS-2f signal is given by C¼

S2f 1 ¼ H2 ð¯mÞ S1f i0

ð9Þ

By the normalization of the 2f signal with the 1f signal magnitude, common terms such as laser intensity, signal amplification, lock-in gain, and laser transmission variation can be eliminated. The 1f-normalized WMS-2f signal, C, is a function of laser parameters (i0, a) and gas parameters (P, T, xi) only. The laser parameters can be determined before the measurements; thus, the comparison should be made directly


109

versus the modulation depth at two temperatures (T ¼ 1000 and 1500 K; P ¼ 1 atm, 2% CO2 and 2% CO in air, L ¼ 10 cm) with the spectroscopic parameters listed in HITRAN 2004, as shown in Fig. 3. It can be seen from the figure that a proper modulation depth a ¼ 0.056 cm1 should be chosen. By choosing this value, the 2f peak heights remain relatively larger for both lines.

EXPERIMENTAL SETUP

FIG. 3. Simulation of the normalized WMS-2f peak heights of the selected line pair versus the modulation depth a, P ¼ 1 atm, 2% CO2 and 2% CO in air, L ¼ 10 cm.

between the WMS-2f simulations and measurements, eliminating the need for scaling between the two.17

SENSOR DEVELOPMENT Line Selection. Absorption spectra based on HITRAN 2004 near 1.57 lm are computed and used to select a proper line pair for CO2 and CO. The line pair near 6361.250 and 6361.344 cm1 is chosen because of its large absorption strength, proper separation of the two lines, and isolation from interference of neighboring CO2, CO, and H2O transitions. Figure 2 shows the simulated 2f signal of the selected pair using HITRAN 2004 for P ¼ 1 atm, T ¼ 1000 K, L ¼ 10 cm, and 2% CO2, 2% CO, and 10% H2O in air. As shown in this figure, the selected pair is well isolated from neighboring CO2, CO, and H2O transitions. Optimization of Modulation Depth. It is known that the WMS-2f peak height is dependent on the line-shape function, u. The selected lines have different FWHM, and hence their modulation depths a are different at the optimum modulation index m, as shown in Eq. 6. In order to select a proper modulation depth for both lines, we simulate the normalized WMS-2f peak heights of the selected CO2 and CO line pair

The measurements of CO2 and CO concentrations at different temperatures are performed in a heated static cell; the arrangement of the experiment is shown in Fig. 4. A heated static cell (38 cm long with canted wedged (1.58) windows to avoid residual etalon fringes), equipped with a temperature controller (type SKW), is used for measurements at temperatures as high as 1000 K. Four type-K thermocouples with an accuracy of 61% are equally attached to the middle and both ends of the absorption cell to monitor the temperature of the gas. Maximum temperature difference along the path length is of the order of 20 K. The gas pressures are measured by a vacuum pressure gauge with an accuracy of 61%. After the heated cell was evacuated to the order of 104 Torr using a mechanical pump and a molecular pump, CO2, CO gas with 99.99% purity, and N2 were delivered into the cell. Light from a DFB InGaAsP laser (NTT) emitting near 1.57 lm is directed across the heated static cell and focused by a spherical mirror onto an InGaAs detector (New Focus Model 2011). The diode laser is temperature and current controlled (ILX Lightwave LDC-3724) and driven by a triangle ramp summed in an adder with a sine wave to provide the wavelength modulation. The sine signal is generated by a lock-in amplifier (Stanford Research Systems Model SR-830) and is also inputted into another similar lock-in amplifier as the reference signal simultaneously. The detector signal is divided into two parts and demodulated by two lock-in amplifiers to recover the 2f and 1f signals. After the acquisition of the 2f and 1f signals is completed, the laser should be injection-current tuned with a triangle ramp across the selected absorption features to determine the actual CO2 and CO concentrations in the test mixture for comparison.

FIG. 4. Schematic of the experimental setup.

110


FIG. 5. The 2f signal of the selected line pair at different temperatures.

FIG. 7. The 1f-normalized absorption-based WMS-2f peak heights measured in each concentration with T ¼ 500 K; the top graph is CO2 and the bottom is CO.

RESULTS AND DISCUSSION The measurements of CO2 and CO concentrations are performed in the static heated cell with controlled gas mixtures over the temperature range of 300 to 1000 K in increments of 50 K. The respective CO2 and CO scanned WMS-2f line shapes for the selected line pair measured at three different temperatures are shown in Fig. 5. As can be seen from the figure, intensities of these WMS-2f signals vary obviously along with the change of temperature. The measured CO2 and CO direct absorption spectrum in the CO2 and CO–N2 mixture at the experimental condition of T ¼ 500 K is shown in Fig. 6. A Voigt profile20 is used to best fit the line shapes of the measured transition, and the residual is given in the bottom panel. This Voigt fit provides the integrated absorbance, from which the CO2 and CO concentration at the experimental conditions can be calculated. Six concentrations of CO2 and CO gas from 3% to 30% are measured at each temperature. Figure 7 shows the 1fnormalized absorption-based WMS-2f peak heights measured at each concentration with T ¼ 500 K; the top and bottom graphs are CO2 and CO, respectively. It can be seen from the figure that the linearity is good; the correlations of these measured points for CO2 and CO have R2 values of 0.998 and 0.999, respectively. The ratios of the CO2 and CO mole

FIG. 6. Measured CO2 and CO direct absorption spectrum in the CO2 and CO–N2 mixture with P ¼ 1 atm, T ¼ 500 K, and L ¼ 38 cm. A best Voigt fit yields the concentration of CO2 and CO. The residual of the fit is shown in the bottom panel.

fraction inferred from the 1f-normalized WMS-2f sensor (xmeasure) and the mole fraction measured by direct absorption (xactual) are shown in Fig. 8. The standard deviations between the measured and actual CO2 and CO mole fractions are 3.3% and 5% over the full temperature range. Those results confirm the accuracy of the sensor for combustion diagnostics. The error in Fig. 8 primarily comes from uncertainties in pressure measurements by the vacuum pressure gauge (1%), measured spectroscopic data (1%), and error in the base line and lineshape fits (0.5%) in the direct absorption measurements. The CO2 and CO mole fractions from the 1f-normalized WMS-2f system agree with the direct absorption measurements, showing good potential for gas sensing at elevated temperatures. In optical gas sensors the sensitivity limit of the system is very important. Figure 9 shows the WMS-2f signals of CO2 and CO detected by one single wavelength scan of the DFB laser measured in the heated static cell with P ¼ 1 atm, 5.5% CO2 and 4% CO in N2, T ¼ 1000 K, and L ¼ 38 cm. It can be inferred from the graph that the minimum detectable concentrations in air over a path length of 1 m are 250 ppm for CO2 and 200 ppm for CO at 1000 K. The baseline of Fig. 9 is not flat, which may be caused by the RAM effect, background spectrum, and etaloning effects.

FIG. 8. Comparison of the measured CO2 and CO concentrations from the 1fnormalized WMS-2f system with the actual values measured in the static cell at different temperatures.


111

ACKNOWLEDGMENTS This research was funded by the National 863 High Technology Research and Development Program of China under Grant no. 2006AA06Z237 and the French International Programme of Scientific Cooperation. (CNRS/PICS no. 3359).

FIG. 9. The WMS-2f signals of CO2 and CO detected by one single wavelength scan of the DFB laser with P ¼ 1 atm, 5.5% CO2 and 4% CO in N2, T ¼ 1000 K, and L ¼ 38 cm.

CONCLUSION Simultaneous detection of CO2 and CO at elevated temperatures with one single DFB laser is demonstrated. The proper CO2 and CO line pair used in the system is selected based on absorption strength, separation of the two lines, and isolation from interference of neighboring transitions of the major combustion gases. The concentrations of CO2 and CO are inferred from their WMS-1f-normalized absorption-based WMS-2f signal peak heights, so as to remove the transmission variation of the laser due to beam steering, mechanical misalignments, soot, and window fouling. Laboratory experiments in a static cell are conducted for the temperature range of 300 to 1000 K (P ¼ 1 atm). The CO2 and CO concentration measurements are within 3.3% and 5% of the expected values over the full temperature range. These good agreements confirm the potential utility for application to gas sensing at elevated temperatures.

112


1. M. G. Allen, Meas. Sci. Technol. 9, 545 (1998). 2. R. K. Hanson and J. B. Jeffries, 25th AIAA Aerodynamic Measurement Technology and Ground Testing Conference, American Institute of Aeronautics and Astronautics, AIAA, 2006-3441 (San Francisco, CA, June 5–8, 2006). 3. L. Joly and B. Parvitte, V. Zeninari and G. Durry, Appl. Phys. B 86, 743 (2007). 4. J. B. Jeffries, C. Schulz, D. W. Mattison, M. A. Oehlschlaeger, W. G. Bessler, T. Lee, D. F. Davidson, and R. K. Hanson, Proc. Combust. Inst. 30, 1591 (2005). 5. J. Wang, M. Maiorov, J. B. Jeffries, D. Z. Garbuzov, J. C. Connolly, and R. K. Hanson, Meas. Sci. Technol. 11, 1576 (2000). 6. R. M. Mihalcea, D. S. Baer, and R. K. Hanson, Appl. Opt. 36, 8745 (1997). 7. A. Farooq, J. B. Jeffries, and R. K. Hanson, Appl. Phys. B 90, 619 (2008). 8. V. Weldon, J. O’Gorman, P. Phelan, J. Hegarty, and T. Tanbun-Ek, Sens. Actuators B 29, 101 (1995). 9. M. E. Webber, J. Wang, S. T. Sanders, D. S. Baer, and R. K. Hanson, Proc. Combust. Inst. 28, 407 (2000). 10. L. S. Rothman, D. Jacquemart, A. Barbe, D. C. Benner, M. Birk, L. R. Brown, M. R. Carleer, Jr., C. Chackerian, K. Chance, L. H. Coudert, V. Dana, V. M. Devi, J.-M. Flaud, R. R. Gamache, A. Goldman, J.-M. Hartmann, K. W. Jucks, A. G. Maki, J.-Y. Mandin, S. T. Massie, J. Orphal, A. Perrin, C. P. Rinsland, M. A. H. Smith, J. Tennyson, R. N. Tolchenov, R. A. Toth, A. J. Vander, P. Varanasi, and G. Wagner, J. Quant. Spectrosc. Radiat. Transfer 96, 139 (2005). 11. J. A. Silver, Appl. Opt. 31, 707 (1992). 12. J. Reid and D. Labrie, Appl. Phys. B 26, 203 (1981). 13. M. Gabrysch, C. Corsi, F. S. Pavone, and M. Inguscio, Appl. Phys. B 65, 75 (1997). 14. R. Engelbrecht, Spectrochim. Acta, Part A 60, 3291 (2004). 15. J. Wang, S. T. Sanders, J. B. Jeffries, and R. K. Hanson, Appl. Phys. B 72, 865 (2001). 16. V. Weldon, J. O’Gorman, J. J. Perez-Camacho, D. McDonald, J. Hegarty, J. C. Connolly, N. A. Morris, R. U. Martinelli, and J. H. Abeles, Infrared Phys. Technol. 38, 325 (1997). 17. H. Li, G. B. Rieker, X. Liu, J. B. Jeffries, and R. K. Hanson, Appl. Opt. 45, 1052 (2006). 18. J. Reid and D. Labrie, Appl. Phys. B 26, 203 (1981). 19. J. T. C. Liu, J. B. Jeffries, and R. K. Hanson, Appl. Phys. B 78, 503 (2004) 20. E. E. Whiting, J. Quant. Spectrosc. Radiat. Transfer 16, 611 (1976).