Acetylene measurement in flames by chirp-based quantum cascade laser spectrometry

Acetylene measurement in flames by chirp-basedquantum cascade laser spectrometry

Zachary R. Quine* and Kevin L. McNesbyWeapons and Materials Research Directorate, U.S. Army Research Laboratory,AMSRD-ARL-WM-BD, Aberdeen Proving Ground, Maryland 21005-5069, USA

*Corresponding author: [email protected]

Received 30 October 2008; revised 24 April 2009; accepted 1 May 2009;posted 5 May 2009 (Doc. ID 103369); published 26 May 2009

We have designed and characterized a mid-IR spectrometer built around a pulsed distributed-feedbackquantum cascade laser using the characteristic frequency down-chirp to scan through the spectral region6:5 cm−1 spectral region. The behavior of this chirp is extensively measured. The accuracy and detectionlimits of the system as an absorption spectrometer are demonstrated first by measuring spectra of acet-ylene through a single pass 16 cm absorption cell in real time at low concentrations and atmosphericpressure. The smallest detectable peak is measured to be ∼1:5 × 10−4 absorbance units, yielding aminimum detectable concentration length product of 2.4 parts per million meter at standard tempera-ture and pressure. This system is then used to detect acetylene within an ethylene–air opposed flowflame. Measurements of acetylene content as a function of height above the fuel source are presented,as well as measurements of acetylene produced in fuel breakdown as a function of preinjection fueltemperature. © 2009 Optical Society of America

OCIS codes: 140.5965, 300.1030, 300.6260, 300.6340.

1. Introduction

Soot formation in gas turbines leads to intense radia-tive heat transfer and mechanical degradation,damaging components and limiting the lifetimeand efficiency of these engines. If the soot is not con-sumed in the combustion process and escapes theturbine it also poses a serious environmental andhealth hazard.Common jet fuel is an unleaded/kerosene oil mix-

ture. The sooting properties of these fuel mixturesvary drastically based on fuel–oxidizer mixture ratio,mixing action, flame confinement, and other factors[1]. In an effort to understand the process of soot for-mation in engines and investigate potential counter-measures to minimize soot emission, fundamentalstudies of flames are under way using different bur-ner configurations to analyze the various stages ofcombustion that take place in a turbine engine [2].

Soot formation begins with the pyrolysis of the fuelinto small, reactive molecules, followed by gas phasepolymerization reactions building up large polycyclicaromatic hydrocarbons [3,4]. These polymerizationreactions continue until the resultant precursor mo-lecules are massive enough to become nucleationsites for small particles, which grow and combinethrough surface reactions to micrometer-scale sootparticles. Soot formation and growth is a complexcompetition between the rates of fuel pyrolysis andproduction of soot precursor molecules and the rateof oxidative attack on these large hydrocarbons.Acetylene plays a crucial role in several of these poly-merization reactions to form soot precursor mole-cules, primarily combining with methylene to formpropargyl (C3H3), a primary building block of thepolycyclic aromatic hydrocarbons [1,2]. Measuringthe acetylene produced in the flame provides ametric for monitoring the soot production.

Quantifying trace gas species concentration withina flame by laser absorption is a nontrivial measure-ment. The instrument must be sensitive and selec-tive to distinguish weak signals from the target

0003-6935/09/163075-09$15.00/0© 2009 Optical Society of America

1 June 2009 / Vol. 48, No. 16 / APPLIED OPTICS 3075

molecule from themyriad other species produced in acombustion reaction. The inherent difficulty of mea-suring absorption spectra at high temperatures,where the population of initial states is spread overa much greater number of accessible states, is madeeven more difficult by measuring through turbulentflows of mixed gases surrounded by scattering andabsorbing soot particles. To perform this measure-ment we built and characterized a sensitive, selectiveinfrared absorption spectrometer system capable ofmeasuring, in real time, absolute acetylene concen-tration in low concentration samples at an elevatedtemperature. This system is designed around apulsed distributed feedback quantum cascade laser(QCL). Recent work shows the QCL to be an extre-mely useful tool for tunable diode laser absorptionspectroscopy (TDLAS) [5–11]. The QCL operatesnear room temperature and provides a powerful(∼10mW), stable, single-mode, mid-infrared lightsource suitable for tunable laser spectroscopy. Nearlythe entire IR spectrum is accessible to QCLs, asthe laser emission is determined by the growth ofthe substrate interstitial layer spacing rather thanthe composition, and a wide spectral range is acces-sible to a single QCL by temperature tuning thesubstrate [12–14]. The QCL used in this experimentis designed for pulsed, single longitudinal modeemission over a thermally tunable range of1279 –1272 cm−1.Working with pulsed diode lasers, a frequency

down-chirp is inflicted on the laser output as a resultof resistive heating as the current pulse deposits en-ergy into the diode chip. There are two methods forworking with this frequency chirp: the effect is trea-ted as an error and minimized or the effect can becharacterized and utilized as a frequency tuning me-chanism. Driving the QCL with short pulses (1–5ns)results in near-Fourier-limited laser pulses that arescanned through the spectral range of interest bytemperature tuning the diode substrate [5] or usinga subthreshold current ramp [9–11]. Typical resolu-tion of δv < 0:005 cm−1 is attainable with this techni-que [5,9]. By harnessing the near linearity of thefrequency down-chirp it is possible to scan througha spectral region in a single long laser pulse(100ns to several microseconds). The resolution ofthis technique is limited by the scan rate ofδv ¼ ðC � dv=dtÞ1=2, where C is a form factor depen-dent on the pulse shape. For long square pulses C ¼0:883 [14]. This latter technique is sometimes calledintrapulse spectroscopy, direct laser absorption spec-troscopy, or long-pulse spectroscopy. Both techniquesyield similar resolution. Short-pulse spectroscopycan scan much longer spectral ranges but requirescomplex computer control of the driving current sup-ply and long time scales for signal collection as thelaser is scanned through the spectral range. Thelong-pulse technique is characteristically simple,yielding spectra similar to a cropped selection of abroadband absorption spectrum; however the maxi-mum spectral range is limited to a fewwavenumbers.

The real-time response of the long-pulse techniquemakes it attractive for studies of flame species con-centration. The turbulent gas flow, steeply varyingtemperature and density profile of the flowing gases,and onset of scattering soot particles all give rise tosignificant random fluctuations in the transmittedintensity. These variations yield line distortionsand false peaks in the absorption spectrum if thetime for scanning a spectral line is comparable tothese environmental variations. The microsecondcollection time of the long-pulse technique allowsus to analyze a frozen flame. Driving the QCL withlong pulses (t > 6 μs) we are able to scan the laserthrough large frequency ranges (from 1278.5 to1272:25 cm−1). This broad spectral range allows usto measure several absorption lines from acetyleneas well as neighboring species such as water ormethane. To our knowledge this is the first experi-ment to take in situ measurements of absorptionspectra within a flame with this long-pulsetechnique.

2. Experiment

The distributed-feedback QCL used in this experi-ment (developed at Alpes Laser, Switzerland,supplied by Boston Electronics, Brookline,Massachusetts) is designed for pulsed, single longi-tudinal mode emission at 7.86 μm. The QCL sub-strate temperature and the driving pulse (currentamplitude, pulse length, and frequency) are con-trolled by a laptop running LabVIEW control VI(Cascade Technologies, Scotland, UK). The QCL ismounted on a Peltier thermoelectric cooler, whichcan vary the substrate temperature from −30 °C to30 °C stabilized to 0:01 °C. The output is collimatedthrough ZnSe optics housed inside the sealed laserhead and exits the case in a roughly collimated beamwith a waist of ∼1mm. A schematic of the experi-mental setup is presented in Fig. 1. For the experi-ments reported here, the laser beam is directedthrough a single-pass 16 cm path length gas absorp-tion cell with wedged BaF2 windows for acetylenevapor measurements, then through an opposed flowburner for measurements of acetylene produced in aflame. The transmitted light is measured by a fast-rising, Peltier-cooled (HgCdZn)Te detector (PVI-2TE-10, VIGO Systems, Warsaw, Poland) and recordedusing a high-speed signal averager from BostonElectronics. The detectivity of the photodetector isD� ¼ 2 × 109 cmHz1=2=W, and the rise time is under0:3ns.

To characterize the laser output wavelength, thebeam is sent through a fixed-mirror Michelson inter-ferometer, built to dynamically measure the changein frequency of the laser output over the course of thepulse. The light exiting the interferometer is mea-sured on the VIGO detector. The wavelength range(Δv, expressed in wavenembers, cm−1) betweentwo maxima measured by the interferometer is aconstant function of the geometry of the light path:Δv ¼ ð2ΔLÞ−1, where ΔL is the path-length

3076 APPLIED OPTICS / Vol. 48, No. 16 / 1 June 2009

difference between the two legs of the interferometer(see Fig. 1). Measurements with this interferometercoupled with absorption spectra from acetylene,water, and methane provide a solid calibration frompulse duration to absolute frequency.Representative examples of laser output versus

time are shown in Fig. 2: (A) the uninterrupted laserpulse, (B) the Michelson interferometer fringes (thebroadening of the fringes is an excellent representa-tion of the nonlinearity of the chirp), (C) the absorp-tion spectrum of 2.8 Torr acetylene in the single-passcell, and (D) a line-of-sight absorption spectrum of anethylene fueled flame. Also shown in A) of Fig. 2 isthe time varying current pulse used to drive the la-ser. The sharp onset and near-constant amplitude ofthe top-hat current pulse leads to abrupt lasing andnearly linear frequency down-chirp. There are smallreflections at the beginning of the pulse that are dueto imperfect impedance matching in the cables thatdeliver the driving signal to the laser head; these re-flections are typical of this type of QCL system [6–8].The fact that the laser output intensity is not con-stant over the breadth of the pulse does not affectthe absorption measurements, as it is a consistentfeature of the laser at a given set of control para-meters. In (B) of Fig. 2 the fringe spacing of theMichelson interferometer is 1:06ð2Þ cm−1 and thetotal usable spectral range of the pulse is approxi-mately 6:5 cm−1.

A. Characterizing the Laser Output

To exploit the heat-induced chirp of the QCL as a fre-quency tuning mechanism it is necessary to fullycharacterize the temporal and spectral evolution ofthe laser output. The output of this QCL is governed

by four controllable parameters: the bulk laser sub-strate temperature, and the current amplitude, timeduration, and repetition rate of the driving pulse.Each affects the chirp rate by controlling the heatdumped into the diode chip and the instantaneoustemperature of the lasing region.

Setting the substrate temperature controls the in-itial frequency of the laser output before the resistiveheating begins to shift the output. The voltage acrossthe chip, or the current amplitude of the pulse, con-trols the current density through the chip and thusthe heating rate from resistive heating, raising thecurrent density and increasing the frequency scanrate. The pulse time duration controls the temporalwidth of the scan and the frequency range. If the dutycycle of the driving pulse is too high, the chip does notcool to the temperature set point between pulses,causing a frequency drift. This is why it is importantto use the pulse repetition rate to keep the duty cycleunder 3%.

To characterize the effect of varying the currentamplitude and duration of the driving pulse, calibra-tion scans were measured on the fixed mirrorMichelson interferometer as these parameters werevaried over the full recommended operating ranges.Plotting the frequency spacing between peaks of theMichelson interference fringes versus the measuredtime between these points in the scan gives a directcalibration from pulse time to laser frequency. Theduty cycle was maintained below 3% to ensure thechip could fully cool between pulses. The damagethreshold for the laser quoted by the supplier wasI ≥ 4:0A. With the driving current amplitude of3:48A the thermoelectric cooler could not keep thesubstrate temperature at the set point and the laser

Fig. 1. Schematic diagram of the direct absorption spectrometer setup. Current supply, temperature controller, and data collection arecontrolled in the LabVIEW instrument panels on a laptop. The single-pass absorption cell is 16:0 cm in length. The laser is sent throughthe burner parallel to the stagnation plane through the central axis of the flame.


output drifted in frequency over the course of min-utes. To avoid damaging the chip our data were takenat lower current amplitudes of 2:57A, which pro-vided smooth, reproducible chirp behavior and nomeasured long-term drift. The interference maximaof the Michelson interferometer are plotted versusthe current pulse time in Fig. 3. Also shown in thisplot are positions of several absorption lines fromacetylene, water, and methane, which were used toaid in the absolute frequency calibration. The fre-quency down-chirp could be treated as linear forthe first 1000ns of the laser pulse, but for pulsesof the length used in this experiment the nonlinear-ity of the chirp must be taken into account.

B. Measuring Acetylene in a Gas Cell

Absorption spectra of acetylene vapor in a 16 cm sin-gle-pass cell were measured to test the accuracy andsensitivity of the spectrometer by comparing themeasured line intensities to the well-characterizedstandard for the acetylene absorption cross sectionas reported in the HITRAN database [15–17].The absorption band of the (v4 þ v5) compound

bending vibration of acetylene is centered near1330 cm−1, the rotational constant of acetylene is

1:125 cm−1. The higher J-value transitions of theP-band are relatively unobscured by absorption fromatmospheric gases and fuel products (e.g., CO2 andC2H4). The Pð23Þ rotational line centered at

Fig. 2. (Color online) (A) Uninterrupted laser pulse profile in blue on top with the current pulse amplitude in red. (B) Laser pulse with thesuperimposed interference fringes of the Michelson interferogram. (C) Transmission spectrum of 2.8 Torr of acetylene in a 16 cm single-pass cell. (D) Transmission spectrum of ethylene/air opposed flow flame taken near the peak acetylene position. All the laser pulse para-meters, 6500ns long, driven at 16 V across the chip with a repetition rate of 1.0 kHz and a substrate temperature set point of −30 °C arerepresentative of the laser settings during data collection.

Fig. 3. (Color online) Calibration curve for laser output as a func-tion of current pulse time. The periodic fringes (circles) are the zerocrossing points of the Michelson interferogram. Gas absorptionlines from methane and acetylene were taken in the absorptioncell; the hot water lines were collected from the flame studies.


1275:512 cm−1 is the main feature measured in thisexperiment as it is near the peak of the P-branch atthe elevated temperatures that will be encounteredwhen probing flames. With a spectral range from1272.2 to 1278.5 wavenumbers, we observe two othertransitions from the (v4 þ v5) vibration of acetyleneas well as several smaller absorptions fromacetylene.The transmitted laser intensity was recorded on

the photodiode through varying partial pressuresof acetylene gas held in the 16 cm single-pass absorp-tion cell as shown in Fig. 1. A typical absorption spec-trum measured in this way is shown in (C) of Fig. 2,measuring absorption from 2.8 Torr acetylene vaporboth neat and diluted to 1 atm in dry nitrogen. Thestretching of the absorption dips toward the end ofthe pulse is another example of the nonlinearity ofthe frequency down-chirp. These transmission spec-tra are converted to spectral absorbance and plottedversus a calibrated frequency scale in Fig. 4.These absorbance spectra were analyzed in Sigma-

Plot to extract the line strengths from the data. Thespectra are taken in the limit of the Beer–Lambertapproximation, where the absorbance is linearly re-lated to the concentration of absorbers [X] and opti-cal path-length L by the absorption cross section σðvÞ:

AðνÞ ¼ − ln�IðνÞIoðνÞ

�¼ σðνÞ½X �L ¼ SgðνÞ½X �L: ð1Þ

In the final relation SgðvÞ is the line strength of theabsorption feature multiplied by a normalized peakfunction. These spectra are fit to the Lorentzian line-shape function

FðvÞ ¼ Σif iðvÞ ¼ Σ

i

Aiπ

� γiγ2i − ðv − voiÞ

�; ð2Þ

where v is the frequency in cm−1, the sum runs overthe indexed peaks centered at voi, and with the spe-cies-specific peak width γi. The peak positions arefixed parameters, and all the peaks from the samegas species share one width. The integrated absor-bance of each peak, Ai, contains the line strength[cm−1 /(molecules cm−2)], the concentration (mole-cules cm−2), and the absorption length (cm):

AC2H2¼

Zf C2H2

ðvÞdv ¼ SC2H2½X�L: ð3Þ

The normalized Lorentzian peak function is chosenover the Voigt peak that is conventionally used in la-ser absorption spectroscopy because both returnequivalent fits to the pressure-broadened absorbancepeaks. There is also evidence that the Voigt profile isno longer theoretically appropriate, a representationof the absorption features in fast frequency-scannedspectra due to transition saturation and other non-linear effects [18,19]. Figure 4 shows a representa-tive absorption spectrum of a 2.8 Torr sample ofpure acetylene diluted to 1atm total pressure indry nitrogen fit to this multi-Lorentzian functionwith residuals.

Plotting the integrated absorbance in Eq. (3) fromthe Pð23Þ line versus the product of the optical lengthand the acetylene concentration (calculated from thepartial pressure of gas in the cell) gives a measureof the line strength parameter that can be comparedto the value listed in the HITRAN database, S ¼2:218 × 10−20 cm−1 /(molecules cm−2) [15]. The inte-grated absorbance is plotted versus acetyleneconcentration and pressure in Fig. 5, showing a lin-ear relationship in fairly good agreement with thepredicted absorbance. The scatter about the pre-dicted line is larger than explained by the qualityof the fit or the standard deviation of repeated mea-surements of a single sample at fixed pressure. Themost likely cause of this scatter is imprecise mea-surement of sample pressure, yielding incorrect pre-dicted concentrations. The gas delivery system thatwas used to fill the absorption cell had leaks thatcould not be fully sealed in the course of the experi-ment, and assigned pressures of the samples could beoff by as much as 10%. A linear, least-squares fit ofthe data, using an absorption cell length of 16 cm,gives a line strength of S ¼ 2:36ð�0:25Þ × 10−20 cm−1/(molecules cm−2), in agreement with the acceptedvalue. Based on the RMS noise in the baseline ofthe spectrum we should be able to accurately mea-sure absorption features with peak heights of 1:5 ×10−4 absorbance units, corresponding to an acetyleneconcentration×length product of 2.4 parts per millionmeter (ppmm) at standard temperature and pres-sure (STP).

Fig. 4. (Color online) Representative absorption spectrum from2.8 Torr of acetylene diluted to 1 atm in dry nitrogen. The spec-trum has been plotted versus the calibrated frequency scale andfit to the multi-Lorentzian peak function discussed in Eq. (2).The results and residuals of this fit are shown in red.


C. Measuring Acetylene in a Flame

As mentioned earlier, the line strength confirmed bythe cell measurements above will not apply to mea-surements in a flame. The line strength is dependenton the population difference between the two levelsof the specified transition. As the temperature in-creases so does the number of accessible initialstates, in accordance with classical Boltzmann statis-tics. As shown in Table 1, the line strength initiallyincreases with temperature as the band center shiftsto higher J values, peaking at approximately 500K;but at the peak flame temperature of an acetylene–air flame (the constant pressure, adiabatic flametemperature of acetylene stoichiometrically mixed

in air is T ¼ 2540K [20]), the line strength has fallento less than 0.3% of its reference value [15].

A line-of-sight absorption spectrum of an acetylenefueled, nonpremixed flame from a glass blower torchis presented in Fig. 6 to emphasize the spectral fea-tures of interest: the three principal acetylene ab-sorption lines and the six water lines (from 12transitions from various levels) that become visibleat these higher temperatures. An ethylene flamefrom this source is difficult to study, as the acetyleneproduced in combustion is spread throughout a large,nonuniform, transient flame region. The characteri-zation of these flames with this QCL system is pos-sible and will provide interesting work in the future,but a less complex flame is more appropriate forinitial measurements with this new system.

The opposed flow burner is a standard burner con-figuration that is well characterized and frequentlyused to characterize the combustion process of fuels.The opposed flow burner directs parallel columns offuel and oxidizer gas into each other out of two noz-zles separated by a narrow gap. In the middle of thisgap the two gases meet, and the momentum in theflow direction goes to zero. At this height, knownas the stagnation plane, the gases begin to mix pro-ducing a quasi-one-dimensional combustion regionas a function of height over the burner. When the fueland oxidizer combine to a stoichiometric combustionpopulation, the gases react and burn in a highly lo-calized flame resembling a coin. A more complete de-scription of the burner used in this experiment can befound in Ref. [2]. Line-of-sight absorption measure-ments were taken as a function of height over the fuelburner. Representative absorption spectra from sev-eral heights over the burner are shown in Fig. 7.These spectra are fit with the same Lorentzian peakshapes from Eq. (2), with species-specific (acetyleneand water) linewidths. The integrated intensity ofthe three principal acetylene lines from the(v4 þ v5) vibration of acetylene are used in the calcu-lation of the concentration. The relative intensities ofmultiple water lines are used to calculate the tem-perature along the line of sight by two-line thermo-metry as outlined in Ref. [21]. We use the ratio ofintensities of the water lines because above1500 °C the acetylene line strength becomes less sen-sitive to variations in temperature. The absorptionmeasurements are converted to concentration usingthe measured width of the flame region in the burneras the absorption length and the line strengths givenin HITRAN at the temperature calculated from therelative intensities of the water lines. The acetyleneconcentration as a function of height over the burneris shown in Fig. 8; tracking the acetylene as afunction of height we see the rise and fall of acetyleneproduced in the decomposition of ethylene and thenconsumed in the combustion reactions.

The effects of preheating the ethylene fuel before itis injected into the burner were measured by thesame technique as above. Measurements show thatpreheating the fuel leads to early decomposition and

Fig. 5. (Color online) Plotting integrated absorbance of Pð23Þat 1275:5 cm−1 acetylene absorption versus concentration of acet-ylene vapor (derived from partial pressure on the top axis). In theBeer–Lambert approximation the slope of this line is a measure ofthe absorption cross section of the material multiplied by the pathlength. The measured cross section is S ¼ 2:36ð�0:18Þ×10−20 cm−1/(molecules cm−2)

Table 1. Temperature Dependence of the Line Strengtha

of the Pð23Þ Absorption Line of the (v4þv5) CompoundBending Vibration of C2H2

Temperature Line Strength (/ 10−20)(K) (cm−1 /molecule cm−2)

296 2.2180400 2.9598470 2.9624600 2.4464800 1.46871000 0.813262000 0.0492903000 0.0069049

aAs predicted in the HITRAN database.


Fig. 7. (Color online) Representative line-of-sight absorption spectra through an ethylene/air opposed flow flame.

Fig. 6. Line-of-sight absorption spectrum from an acetylene diffusion mixed flame collected over 4096 pulses of the laser. The principalacetylene absorption lines and the lines belonging to water are labeled. All unlabeled peaks have been assigned to acetylene absorptions.


increases the acetylene produced before combustion.This increase in acetylene concentration leads to in-creased soot production, which is also observablevisibly as the flame becomes brighter (due to soot in-candescence). Figure 9 shows the peak acetylene con-centration as a function of ethylene fuel temperaturebefore injection to the burner. The measured acety-lene concentration increases by more than 25% whenthe ethylene fuel is preheated. These measurementsdemonstrate the capability of our system to detectand quantify changes in acetylene produced in com-bustion due to changes in the composition of theflame.

3. Conclusion

We have constructed, characterized, and tested anovel spectrometer that can deliver real-time high-resolution IR absorption spectra. This system is ea-sily tuned over a relatively broad spectral range thatincludes several absorption features from multiplegas species of interest to combustion studies. Havingproved the accuracy of this spectrometer in measur-ing the concentration of samples of acetylene vaporat atmospheric pressure, we successfully used it tomeasure acetylene produced within a flame. Thesmallest detectable peak is measured to be ∼1:5 ×10−4 absorbance units under ideal conditions, yield-ing aminimum detectable concentration of 2:4ppmmat STP, however, background variations in the flametend to decrease sensitivity. Further work is plannedto measure the detailed spatial profile of acetyleneconcentration to allow for proper correction of line-of-sight data to three-dimensional concentrationmaps within more complex flames. We also intendto use this system to characterize the acetylene pro-duction in a series of flames from a variety of fuelsand mixing actions.

This work was funded by the Strategic Environ-mental Research and Development Program(SERDP).

References1. R. Flagan and J. Seinfeld, “Particle formation in combustion,”

in Fundamentals of Air Pollution Engineering (Prentice–Hall,1988), pp. 358–390.

2. K. McNesby, A. W. Miziolek, T. Nguyen, F. C. Delucia, R. R.Skaggs, and T. A. Litzinger, “Experimental and computationalstudies of oxidizer and fuel side addition of ethanol to opposedflow air/ethylene flames,” Combust. Flame 142, 413–427 (2005).

3. S. J. Harris and A. M.Weiner, “Surface growth of soot particlesin premixed ethylene/air flames,” Combust. Sci. Technol. 31,155–167 (1983).

4. S. J. Harris and A. M. Weiner, “Determination of the rateconstant for soot surface growth,” Combust. Sci. Technol.32, 267–275 (1983).

5. A. A. Kosterev and F. Tittel, “Chemical sensors based onquantum cascade lasers,” IEEE J. Quantum Electron. 38,582–591 (2002).

6. E. Normand, G. Duxbury, and N. Langford, “Characterisationof the spectral behavior of pulsed quantum cascade lasersusing a high resolution Fourier transform infrared spec-trometer,” Opt. Commun. 197, 115–120 (2001).

7. E. Normand, M. McCulloch, G. Duxbury, and N. Langford,“Fast, real-time spectrometer based on a pulsed quantum cas-cade laser,” Opt. Lett. 28, 16–18 (2003).

8. T. Beyer, M. Braun, and A. Lambrecht, “Fast gas spectroscopyusing pulsed quantum cascade lasers,” J. Appl. Phys. 93,3158–3160 (2003).

9. K. Namjou, S. Cai, E. A. Whittaker, J. Faist, C. Gmachl,F. Capasso, D. L. Sivco, and A. Y. Cho, “Sensitive absorptionspectroscopy with a room temperature distributed-feedbackquantum-cascade laser,” Opt. Lett. 23, 219–221 (1998).

10. D.Weidmann, A. A. Kosterev, C. Roller, R. F. Curl,M. P. Fraser,and F. K. Tittel, “Monitoring of ethylene by a pulsed quantumcascade laser,” Appl. Opt. 43, 3329–3334 (2004).

11. A. A. Kosterev, F. Tittel, C. Gmachl, F. Capasso, D. L. Sivco, J.N. Baillargeon, A. L. Hutchinson, and A. Y. Cho, “Trace-gas

Fig. 8. Acetylene concentration measured through the centralaxis of the opposed flow burner as a function of height over thefuel source burner. The rise and fall of the acetylene concentrationis below the stagnation plane and the observed location of the in-candescent flame region.

Fig. 9. Peak acetylene concentration plotted versus preinjectiontemperature of the ethylene fuel. Measured as the ethylene cooledfor several minutes after the heaters that prewarmed the fuel wereshut off. The increased acetylene concentration could indicate ear-lier decomposition of the hot fuel.


detection in ambient air with a thermoelectrically cooledpulsed quantum-cascade distributed feedback laser,” Appl.Opt. 39, 6866–6872 (2000).

12. “Alpes Lasers Literature,” http://www.alpeslasers.ch/Publications.html.

13. A. Müller, M. Beck, J. Faist, U. Oesterle, and M. Ilegems,“Electrically tunable room-temperature quantum-cascadelasers,” Appl. Phys. Lett. 75, 1509–1511 (1999).

14. T. Beyer, M. Braun, S. Hartwig, and S. Lambrechtin,“Linewidth measurement of free-running, pulsed, distributedfeedback quantum cascade lasers,” J. Appl. Phys. 95, 4551–4554 (2004).

15. L. S. Rothman,D. Jacquemart, A. Barbe, D. C. Benner,M.Birk,L. R. Brown, M. R. Carleer, C. Chackerian Jr., 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, J.Vander Auwera, P. Varanasi, and G. Wagner, “The HITRAN2004 molecular spectroscopic database,” J. Quant. Spectrosc.Radiat. Transfer 96, 139–204 (2005).

16. D. Jacquemart, J. -Y. Mandin, V. Dana, C. Claveau, J. VanderAuwera, M. Herman, L. S. Rothman, L. Régalia-Jarlot, andA. Barbe, “The IR acetylene spectrum in HITRAN: updateand new results,” J. Quant. Spectrosc. Radiat. Transfer 82,363–382 (2003).

17. Y. Kabbadj, M. Herman, G. Di Lonardo, L. Fusina, and J. W. C.Johns, “The bending energy levels of C2H2,” J. Mol. Spectrosc.150, 535–565 (1991).

18. M. T. McCulloch, G. Duxbury, and N. Langford, “Observationof saturation and rapid passage signals in the 10.25 micronspectrum of ethylene using a frequency chirped quantumcascade laser,” Mol. Phys. 104, 2767–2779 (2006).

19. G. Duxbury, M. Langford, and M. McCulloch, “Rapid passageinduced population transfer and coherences in the 8 micronspectrum of nitrous oxide,” Mol. Phys. 105, 741–754 (2007).

20. C. E. Baukal, Jr., and R. E. Schwartz, The John ZinkCombustion Handbook (CRC Press, 2001).

21. R. G. Daniel, K. L. McNesby, and A. W. Miziolek, “Applicationsof tunable diode laser diagnostics for temperature and speciesconcentration profiles of inhibited low-pressure flames, Appl.Opt. 35, 4018–4025 (1996).


Acetylene measurement in flames by chirp-based quantum cascade laser spectrometry

Documents

Transcript of Acetylene measurement in flames by chirp-based quantum cascade laser spectrometry