Research Article
Applied Optics
1
A single-ended mid-infrared laser-absorption sensor for simultaneous in-situ measurements of H2 O, CO2 , CO, and temperature in combustion flows W EN Y U P ENG1,* , C HRISTOPHER S. G OLDENSTEIN2,† , R. M ITCHELL S PEARRIN3,† , J AY B. J EFFRIES1 , AND R ONALD K. H ANSON 1 1 High
Temperature Gasdynamics Laboratory, 452 Escondido Mall, Bldg. 520, Thermosciences Division, Stanford University, CA 94305, USA of Mechanical Engineering, Purdue University, West Lafayette, IN 47907, USA 3 Department of Mechanical and Aerospace Engineering, University of California, Los Angeles, CA 90095, USA * Corresponding author:
[email protected] † Postdoctoral scholars at Stanford University during the completion of this work. 2 School
Compiled October 15, 2016
The development and demonstration of a four-color single-ended mid-infrared tunable laser-absorption sensor for simultaneous measurements of H2 O, CO2 , CO, and temperature in combustion flows is described. This sensor operates by transmitting laser light through a single optical port and measuring the backscattered radiation from within the combustion device. Scanned-wavelength-modulation spectroscopy with second-harmonic detection and first-harmonic normalization (scanned-WMS-2f /1f ) was used to account for variable signal collection and non-absorption losses in the harsh environment. Two tunable diode-lasers (TDLs) operating near 2551 and 2482 nm were utilized to measure H2 O concentration and temperature while an interband-cascade laser (ICL) near 4176 nm and a quantum-cascade laser (QCL) near 4865 nm were used for measuring CO2 and CO, respectively. The lasers were modulated at either 90 kHz or 112 kHz and scanned across the peaks of their respective absorption features at 1 kHz, leading to a measurement rate of 2 kHz. A hybrid demultiplexing strategy involving both spectral filtering and frequency-domain demodulation was used to decouple the backscattered radiation into its constituent signals. Demonstration measurements were made in the exhaust of a laboratory-scale laminar methane-air flat flame burner at atmospheric pressure and equivalence ratios ranging from 0.7 to 1.2. A stainless steel reflective plate was placed 0.78 cm away from the sensor head within the combustion exhaust, leading to a total absorption path length of 1.56 cm. Detection limits of 1.4% H2 O, 0.6% CO2 , and 0.4% CO by mole were reported. To the authors’ knowledge, this work represents the first demonstration of a mid-infrared laser-absorption sensor using a single-ended architecture in combustion flows. © 2016 Optical Society of America OCIS codes: (120.1740) Combustion diagnostics; (300.1030) Absorption spectroscopy; (140.5960) Semiconductor lasers; (060.2370) Fiber optics sensor; (300.6340) Spectroscopy, infrared. http://dx.doi.org/10.1364/ao.XX.XXXXXX
1. INTRODUCTION Tunable diode laser-absorption spectroscopy (TDLAS) sensors are widely used to provide non-intrusive, in-situ measurements of gas temperature, concentration, pressure, and velocity for a variety of combustion applications and species [1–17]. Typical TDLAS sensors operate by passing narrowband laser light through an absorbing gas at a frequency corresponding to a specific quantum transition and measuring the transmitted intensity on an opposing detector. The measured absorption is then related to thermodynamic properties of the gas using known
spectroscopic parameters associated with the transition. In many environments of practical interest, however, it can be challenging to install a direct optical line-of-sight (LOS) between light sources and detectors. In such cases, it can be advantageous to employ indirect LOS techniques to probe the environment of interest. Several researchers have developed sensor architectures in which a direct LOS is not needed. Smith et al. [18] developed an infrared laser-absorption probe using sapphire rod waveguides with 45◦ -cleaved ends to simultaneously measure temperature and water vapor concentration in the exhaust of a low pressure
Research Article
flame. However, the invasive nature of this probe architecture limits the accuracy and applicability of this sensor in harsh environments. Other researchers [19–25] have employed a singleended sensing approach. These sensors rely on detecting the backscattered laser light either from a retro-reflector or a diffuse surface within the tested apparatus and therefore require only a single optical port to be installed within the device. This allows for non-intrusive, in-situ measurements of thermodynamic quantities within the devices with significantly fewer modifications required to gain optical access. Most notably, Rein et al. [25] have demonstrated a 100 kHz measurement rate sensor for water concentration and temperature within the combustor annulus of a H2 -air fueled rotating detonation engine (RDE). Additionally, Goldenstein et al. [24] demonstrated stand-off measurements of CH4 and H2 O in the environment at up to 10 meters in path length using fiber-coupled near-infrared (near-IR, defined here as λ < 2 µm) lasers. None of the previous work in the literature, however, has demonstrated the single-ended approach beyond the near-IR. Recent advances in laser technology have enabled spectroscopic access to the strongest fundamental rovibrational bands of common combustion species in the mid-infrared (mid-IR, 2 − 14 µm). By virtue of their strength, these mid-IR transitions can significantly improve the detection limits of or reduce the optical path lengths needed for TDLAS-based sensors. When coupled with similar improvements in mid-IR fiber optics technology, these advances have enabled the development of highly sensitive sensors that remotely deliver multiple light sources onto a single LOS [8]. In this research, we leverage these advances and extend the single-ended sensor architecture to enable simultaneous measurements of multiple species and temperature over short path lengths on the order of 1 cm. Specifically, we present the development and demonstration of a simultaneous H2 O, CO2 , CO, and temperature single-ended sensor with a 2 kHz measurement rate for combustion flows using four fiber-coupled mid-IR lasers. Two tunable diode lasers (TDLs) operating near 2551 and 2482 nm were used to measure H2 O concentration and temperature while an interband cascade laser (ICL) near 4176 nm and a quantum cascade laser (QCL) near 4865 nm were used to probe CO2 and CO transitions, respectively. Scannedwavelength-modulation spectroscopy with second harmonic detection and first harmonic normalization (scanned-WMS-2 f /1 f ) was used to account for the effects of non-absorbing transmission losses, beam-steering, and 1/ f noise sources in the combustion stream. The methods used to convert the measured WMS-2 f /1 f signals into gas concentration were validated by measuring CO partial pressures within a gas cell at room temperature and comparing the results against known values. The sensor was then demonstrated in a combustion environment by simultaneously measuring the target species and temperature in the exhaust of a methane-air flat-flame burner at atmospheric pressure. Detection limits of 1.4% H2 O, 0.6% CO2 , and 0.4% CO by mole were achieved using the scanned-WMS strategy. To the authors’ knowledge, this work represents the first demonstration of a mid-IR multi-species and temperature sensor using a single-ended back-reflection-based optical scheme in combustion flows. The success of this sensor was dependent on resolving several challenges in optical engineering, including: 1. The use of a hollow-core fiber bundle to remotely deliver and combine light from four mid-IR lasers onto a single line
Applied Optics
2
of sight. 2. Evaluating the effects of reflector material and surface quality on sensor performance. 3. Preventing the detection of unwanted reflection of light from window surfaces prior to entering the combustion exhaust, which can lead to inaccurate measurements. The remainder of this article is organized as follows: Section 2 briefly discusses the theoretical basis of laser-absorption spectroscopy, the scanned-WMS-2 f /1 f technique, and the methods used to convert measured signals into the quantities of interest. Section 3 discusses the sensor design, including the line selection procedure and the optical setup used to achieve singleended measurements. Known difficulties in the sensor design and their solutions are discussed, including a discussion on the choice of reflector material. Section 4 presents the results of the experiments conducted to validate the sensor design and to demonstrate the sensor in a laboratory-scale combustion environment. A discussion on theoretical detection limits of this sensor is also provided. Section 5 summarizes the main results of this work and provides suggestions for additional work and potential applications of this type of single-ended sensor in practical systems.
2. THEORY The sensors used here are based on the theory of laser-absorption spectroscopy (LAS) employing the wavelength modulation spectroscopy (WMS) methodology for signal recovery and interpretation. These concepts are briefly discussed in the next three sections: A. Laser-absorption spectroscopy
The theory of laser-absorption spectroscopy is well documented in the literature and only a brief description is reproduced here to define units and convention. More detailed discussions can be found in other works [26–28]. The fractional transmission, τν , of monochromatic radiation at some frequency ν (cm−1 ) through an absorbing medium is governed by the Beer-Lambert relation: It τν = = exp(−αν ) (1) I0 ν where It is the transmitted intensity, I0 is the incident intensity, and αν is the spectral absorbance, which may be calculated for a uniform ideal gas via the following: αν =
∑ Sij Pχi Lφν,ij
(2)
i,j
where the indices i and j represent the jth transition of the ith species. Here Sij (cm−2 /atm) is the linestrength, P (atm) is the total pressure of the gas, χi is the mole fraction of the ith species, L (cm) is the optical path length, and φν,ij is the lineshape function. Sij is a function of temperature and known parameters and can be calculated as follows: " # hcEij” 1 T0 Qi ( T0 ) 1 Sij ( T ) = Sij ( T0 ) exp − − T Qi ( T ) kB T T0 (3) −1 hcν0,ij hcν0,ij × 1 − exp − 1 − exp − kB T k B T0
Research Article
Applied Optics
3
Here, Sij ( T0 ) is the linestrength (tabulated) at some reference temperature T0 (typically 296 K for most spectroscopic databases), Qi is the total partition function of the absorbing species, Eij” is the lower state energy of the transition, and ν0,ij is the linecenter frequency in vacuum. The constants h (J·s), c (cm/s), and k B (J/K) are Planck’s constant, the speed of light, and Boltzmann’s constant, respectively. The Voigt function was used to model the lineshapes of the transitions, which considers the combined effects of Doppler and collisional broadening on the spectra. These effects are characterized by the Doppler broadening full-width at half-maximum (FWHM), ∆νD (cm−1 ), and the collisional broadening FWHM, ∆νC (cm−1 ), given by r T ∆νd = 7.162 × 10−7 ν0 (4) M ni T0 ∆νc = 2P ∑ χi γi,0 (5) T i
Scanned-wavelength-modulation spectroscopy (scanned-WMS) is a laser-absorption technique that is tolerant to noise and immune to several known difficulties associated with commonly used direct absorption spectroscopy techniques. Another advantage of using WMS is that the light from multiple lasers can be detected on the same detector by modulating the lasers at different frequencies, a technique known as frequency multiplexing and is essential for reducing system complexity. A short description of WMS is provided here—we refer the reader to other works [29–31] for more thorough discussions on WMS theory. In scanned-WMS, the laser optical frequency is temporally ¯ typically near the tuned about some mean optical frequency, ν, transition linecenter of the target species using a combination of a high-frequency sinusoidal modulation at frequency f m and a lower frequency scan at frequency f s . The instantaneous optical frequency is expressed as
where ν0 is the center frequency (cm−1 ) of the transition, M is the molecular mass of the absorbing species (g/mol), γi,0 (cm−1 ·atm−1 ) is the collisional-broadening half-width halfmaximum (HWHM) coefficient for the ith collisional partner at a reference temperature T0 , and ni is the temperature exponent of the broadening coefficients. More information regarding the Voigt function can be found in [28].
where am (cm−1 ) and as (cm−1 ) are the modulation and scan depths, respectively, while φm and φs are temporal phases of the optical frequency tuning. The corresponding non-absorbing laser intensity tuning is expressed as a Fourier series: "
B. Inferring temperature and mole fraction
In tunable laser-absorption spectroscopy, τν is typically measured over a significant portion of the lineshape of a target transition. In such cases, it is useful to define the integrated absorbance, Aij , of a given transition, which is calculated as follows: Z ∞ Aij = αν,ij dν = Sij Pχi L (6) −∞
This quantity removes the dependence on transition lineshape and can be used to robustly calculate gas properties of interest. In two-color thermometry, temperature can be inferred by taking the ratio, R, of two transitions of the same species within the gas: A S Pχ L S R , i1 = i1 i = i1 ( T ) (7) Ai2 Si2 Pχi L Si2 Here, R can be inverted to solve for the temperature using known spectroscopic parameters for the two transitions. For transitions where the vacuum linecenter frequencies are close to each other (as is the case for the two water transitions utilized in this work), a closed-form solution to temperature in terms of R and spectroscopic parameters can be derived [28]: hc ” ” k B Ei2 − Ei1 (8) T= S (T ) ” − E” ln R + ln Si2 (T0 ) + k BhcT0 Ei2 i1 i1
0
Differentiating Eq. 7 with respect to temperature yields a measure of the sensitivity of R to changes in T for a given line pair at a specific temperature: ” ” T dR − Ei2 hc Ei1 σT , = (9) R dT kB T With temperature, pressure, and path length known, the mole fraction can be inferred from the integrated absorbance from either transition as follows: Aij (10) χi = Sij PL
C. Scanned-WMS-2f/1f
ν(t) = ν¯ + am cos(2π f m t + φm ) + as cos(2π f s t + φs )
(11)
∞
I0 (t) = I0 1 +
∑
isk cos (2πk f s t + ψsk ) + imk cos (2πk f m t + ψmk )
k =1
(12) Here, I0 represents the mean intensity, the i represent the meannormalized intensity response amplitudes, and the ψ represent the temporal phases of the intensity waveform. The subscripts s and m correspond to the scan and modulation signals, respectively. For semiconductor lasers controlled by high-compliance current sources, the summation in Eq. 12 is typically truncated to only the two leading terms without significant impact to the reconstructed waveform or to the results of the WMS postprocessing method. The nth harmonics (n f ) of the transmitted intensity waveform relate directly to the laser intensity tuning characteristics and absorption spectra and can be utilized to make measurements of unknown gas properties. These harmonics are extracted from the measured intensity waveforms via digital lock-in filters. The 2nd -harmonic (2f ) signal is typically dominated by features caused by the absorption spectra whereas the 1f signal is typically dominated by the laser intensity tuning. However, since the tuning amplitudes of all harmonics are proportional to the mean transmitted intensity, the 1f signal is used to normalize the 2f measurement signal so that the resulting WMS-2 f /1 f spectra becomes insensitive to fluctuations in the mean transmitted laser intensity. This feature is especially useful in harsh environments where non-absorption losses (e.g. due to beam-steering) can lead to significant fluctuations in the transmitted laser intensity during a measurement. To recover the integrated absorbances of the target transitions, the WMS spectral-fitting technique as described by Goldenstein et al. [29] was used. Here, simulated Voigt profiles and known laser tuning properties were fed into a WMS model to produce the simulated WMS spectrum for a given transition. A Levenberg-Marquardt nonlinear least-squares optimization algorithm [32, 33] was then used to vary the parameters within the simulated Voigt profiles until the mean-squared-error between the simulated and measured WMS spectra reached a minimum. ¯ A, and ∆νc are free parameters in the For each Voigt profile, ν,
#
Research Article
fitting routine. ∆νD is set by temperature, but since temperature is not known a priori, the fitting routines for the two thermometry transitions were iteratively performed until the measured two-color temperature converged. A key feature of this spectralfitting technique is that ∆νc is fitted to the WMS-2 f /1 f spectra, meaning that an accurate database of collisional broadening parameters is not needed. The converged values for the integrated absorbance were used to measure temperature and mole fraction as discussed in Section B.
Applied Optics
4
Table 1. Linecenters (ν0 ), lower state energies (E00 ), and ref-
erence linestrengths (S) of the selected H2 O, CO2 , and CO transitions. Species
ν0 (cm−1 )
E” (cm−1 )
S(296 K) (cm−2 /atm)
H2 O
3920.08
704.2
6.35 × 10−1
H2 O
4029.52
2660.95
1.10 × 10−4
CO2
2394.41
3329.0
7.39 × 10−5
CO
2055.40
886.9
6.20 × 10−1
3. SENSOR DESIGN A. Line selection
Linestrength at 3920.08 cm−1 Linestrength at 4029.52 cm−1 Temperature sensitivity
3
2.5
0.2
2
0.1
1.5
−2
0.3
0 1000
1200
1400 1600 Temperature [K]
1800
Temperature sensitivity
0.4
Linestrength [cm /atm]
Selecting optimal absorption transitions is crucial to the performance of all laser-absorption sensors. The primary criteria for line selection include sufficient absorbance and minimal spectral interference from other species at the target range of thermodynamic conditions. Based on prior experiments performed with similar lasers and detection systems, WMS-based sensors typically require peak absorbances to be between 0.001 and 1.5 to achieve adequate signal-to-noise ratio (SNR) while maintaining sensitivity to species concentration. A key challenge in the selection process is the confined geometries found in many practical combustion systems, where channel widths and therefore optical path lengths on the order of one centimeter is common (e.g. within the high-pressure turbine section of a gas turbine). This complicates the use of near-IR (< 2 µm) overtone transitions of the combustion species, whose transition linestrengths are too weak to provide sufficient absorbance over short path lengths. The recent commercialization of tunable, room-temperature semiconductor lasers operating in the mid-IR have enabled access to the strong fundamental rovibrational bands of the combustion species studied here. Transitions within this region offer orders of magnitude greater absorption and more sensitive detection compared to the next-strongest near-IR absorption bands for each species. Line selection was therefore restricted to the strong mid-IR bands to compensate for the short optical path lengths encountered in practical combustion systems. The target thermodynamic environments studied here are the exhaust of hydrocarbon-air flames with equivalence ratios (φ) ranging from 0.7 to 1.2, pressures ranging from 1 to 3 atm, temperatures ranging from 1000 to 2000 K, and a path length of 1.5 cm. These conditions represent typical conditions found in experimental rotating-detonation engines (RDEs) [34], which is the target application of this single-ended sensor. Several researchers have previously developed robust mid-IR laserabsorption sensors for combustion sensing in harsh environments [8, 35–38]. The mid-IR transitions selected by these researchers have been proven to work for the thermodynamic environments studied in this work; hence, those same transitions were reused for the current single-ended sensor. Table 1 provides select spectroscopic parameters for the transitions of interest. Two transitions near 3920.08 cm−1 (2551 nm, referred to as the “low E” H2 O transition" in this work) and 4029.52 cm−1 (2482 nm, referred to as the “high E” H2 O transition" in this work) were chosen to enable H2 O and temperature measurements. Spectroscopic parameters for these two transitions were experimentally measured by Goldenstein et al. in [39] for the 4029.52 cm−1 line and in [35] for the 3920.08 cm−1 line. These two transitions were previously used by Goldenstein et al. in several works [35–37] for sensitive H2 O and thermometry measurements in supersonic-combustion ramjets (scramjet), pulse-detonation engines (PDE), and rotating-detonation en-
1 2000
Fig. 1. Linestrengths and temperature sensitivity of the two
targeted mid-IR H2 O transitions as a function of temperature.
gines (RDEs). The linestrengths and temperature sensitivity of these two transitions (as defined by Eq. 9) is shown in Fig. 1 for the target range of temperatures between 1000 and 2000 K. As can be seen, the sensitivity remains above 1.4 for the full range of temperatures, indicating that these transitions can be used for sensitive thermometry. Two transitions near 2394.41 cm−1 (4176 nm) and 2055.4 cm−1 (4865 nm) were used to measure CO2 and CO, respectively. The CO2 feature corresponds to the R(92) transition of the its fundamental asymmetric stretch (ν3 ) band while the CO feature corresponds to the P(21) transition of its fundamental stretch band. The R-branch of the fundamental ν3 CO2 band was first suggested by Spearrin et al. [40] for use in combustion diagnostics due to low spectral interference from neighboring H2 O transitions. This particular CO2 transition was used by Spearrin et al. to perform simultaneous CO2 and temperature measurements in a scramjet combustor [8]. Although the CO transition was previously unused for combustion applications, its neighboring P(20) transition near 2059.91 cm−1 was used by Spearrin et al. for CO measurements again in scramjets [8] as well as for high-pressure PDE environments [38]. Since the spectral characteristics of the P(21) transition are largely similar to the P(20) transition, the P(21) transition was used for this work due to the availability of a higher-power laser. Additionally, there is negligible spectral interference from H2 O and CO2 near the CO P(21) linecenter based on HITEMP simulations at typical combustion exhaust conditions (data not presented here). Linestrengths and lower-state energy values for the CO2 and CO transitions, espe-
Research Article
Applied Optics
0.25
0.35 18.0% H2O @ P = 1 atm
18.0% H2O @ P = 1 atm
18.0% H O @ P = 3 atm
18.0% H O @ P = 3 atm
0.3
2
0.2
Low E" H2O
Absorbance [−ln(It/I0)]
Absorbance [−ln(It/I0)]
5
0.15
0.1
2
High E" H2O
0.25 0.2 0.15 0.1
0.05 0.05 0 3919.6
3919.8
3920 3920.2 Frequency [cm−1]
0
3920.4
(a)
4029.4 4029.6 Frequency [cm−1]
4029.8
4030
(b) 0.4
0.7 7.1% CO2 @ P = 1 atm 7.1% CO2 @ P = 3 atm
0.35
3.6% CO @ P = 1 atm 3.6% CO @ P = 3 atm
0.6 CO2
0.3
Absorbance [−ln(It/I0)]
Absorbance [−ln(It/I0)]
4029.2
R(92) 0.25 0.2 0.15 0.1
0.4 0.3 0.2 0.1
0.05 0
CO P(21)
0.5
2394
2394.2
2394.4 2394.6 Frequency [cm−1]
0
2394.8
(c)
2055
2055.2 2055.4 2055.6 Frequency [cm−1]
2055.8
(d)
Fig. 2. Simulated absorption spectra of the (a) low E” H2 O transition, (b) high E” H2 O transition, (c) CO2 R(92) transition of the
ν3 fundamental band, and (d) CO P(21) transition of the fundamental band. Simulations were performed assuming equilibrium concentrations in the exhaust of a φ = 1.2 CH4 -air flame at T = 1500 K, L = 1.5 cm, and P = 1 or 3 atm in a bath gas of N2 . cially in the fundamental bands, are generally well known due to their simple molecular structure [41, 42], and were taken from the HITEMP 2010 database [43]. Sample absorption spectra for all four transitions utilized in this sensor are presented in Fig. 2. Here, equilibrium composition of a CH4 -air flame at φ = 1.2, T = 1500 K, and P = 1 and 3 atm were simulated for a path length of 1.5 cm in a bath gas of N2 . As can be seen, the peak absorbances for all four transitions are within the acceptable regions between 0.001 and 1.5 for sensitive WMS measurements. B. Optical design
Fig. 3 shows a graphical depiction of the optical configuration for the multi-species single-ended laser-absorption sensor. Two free-space distributed-feedback (DFB) diode lasers (Nanoplus GmbH) operating near 3920 cm−1 and 4029 cm−1 , each with approximately 10 mW of output intensity, provided single-mode light sources for H2 O and temperature detection. For CO2 detection, a free-space DFB interband cascade laser (ICL, Nanoplus GmbH) provided approximately 30 mW of output intensity at 2394 cm−1 . For CO measurement, a free-space DFB quantum cascade laser (QCL, Alpes Laser S.A.) provided approximately
Fig. 3. Schematic of the single-ended sensor optical architec-
ture with major optical elements labeled. The red-dotted lines represent the nominal beam path through the sensor assembly.
Research Article
Fig. 4. Computer rendering of the single-ended pitch and
catch assembly with major optical elements labeled.
70 mW of output power at 2055 cm−1 . Light from all four lasers was fiber coupled in free space onto a 4-to-1 multi-mode hollowcore silver-coated glass fiber bundle (Opto-Knowledge Systems, Inc.) using a combination of anti-reflection (AR)-coated BaF2 spherical lenses and kinematic mounts (Thorlabs, Inc.). The laser targeting the lower E” H2 O transition is strongly attenuated by room humidity and so the entire laser assembly was enclosed and purged with dry N2 . Additionally, humidity within the hollow fiber cores was removed by pressurizing the collimation optics housing with dry N2 . Humidity within the collection optics and detector assembly was partially purged with dry N2 , but a significant amount of light from the laser targeting the low E" H2 O transition was still attenuated by the ambient water vapor. The residual absorption due to ambient humidity was compensated for by employing the background subtraction technique and is discussed in detail in Section C. Focal lengths for the lenses were chosen such that the beam waist is optimized for the core diameter (∼ 300 µm) depending on the wavelength of the laser light. Optimal coupling efficiency was achieved by maximizing beam intensity as measured by a power-meter (Ophir Optronics Ltd.) at the exit of the fiber bundle. A coupling efficiency of approximately 20% was achieved for all four lasers, which is sufficient for the detection scheme introduced next. The hollow-core fiber bundle used here was ARcoated on the input side of each fiber to prevent back-reflection into the lasing cavities. Each fiber was combined onto a single SMA-type fiber connector, forming a square array of hollow core terminators at the output of the fiber bundle. At a position 10 cm away from the output of the fiber bundle, the beam centers of each laser were located within a 2 mm diameter circle when measured with a beam locator (Ophir), demonstrating that the hollow cores exhibit good collinearity at the fiber bundle output. We can therefore assume that each laser beam traverses identical beam paths upon exiting the fiber bundle. The output of the fiber bundle was delivered into the pitch/catch assembly, shown as a computer rendering in Fig. 4. A gold-plated right-angle mirror immediately downstream of
Applied Optics
6
the collimation optics turned the beams towards the combustor channel such that the angle between the beam and the surface normal vector of the reflector was approximately 5◦ . This angle was chosen to accommodate the geometric constraints of the window while simultaneously maximizing the amount of light collected by the downstream collection optics and photodetectors. A 1.27 cm long, 0.64 cm wide, 0.30 cm thick rectangular sapphire window (Meller Optics, Inc.) with a 2◦ wedge (to prevent the formation of e´ talon fringes) along the length-wise direction was installed on an application-specific stainless steel window plug, which served as the physical interface between the sensor and combustion gases. The light from all four lasers entered the combustor channel via the sapphire window and allowed to reflect off of an opposing surface within the combustion device. The reflected light was then collected by a 50/50 beam-splitter and collected onto separate photodetectors (Vigo System S.A.) for the H2 O and CO/CO2 transitions. To separate the signals from each laser, a 2.2-2.6 µm bandpass filter and a 4 µm longpass filter were placed in front of the photodetectors detecting laser light targeting the H2 O and CO/CO2 transitions, respectively, thus allowing each photodetector to detect light from only two of the lasers. Note that sapphire naturally attenuates light beyond 5 µm, hence a bandpass filter was unnecessary for the CO/CO2 detector. Frequency-domain multiplexing was used to further isolate individual transitions from each laser pair. In this technique, each laser within each laser pair was modulated at different frequencies. Lock-in filters were then used to de-multiplex each laser during post-processing. More information regarding frequency-domain multiplexing can be found in [44]. The orientation and quality of the surface can have a significant effect on the amount of light collected by the photodetector. In particular, poor surface quality can lead to excessively diffuse scattering of the incident radiation, leading to poor overall collection efficiency. Since the target application of this sensor is for measurements in the combustor annulus of hydrocarbon-fueled RDEs, the relative performance of four types of materials and material surfaces were assessed: unpolished and polished aluminum and stainless steel flat metal plates. These metal plates were each placed 0.76 cm away from the sapphire window in accordance with the target optical path length of this sensor. As expected, the sensor performed poorly when using unpolished metals, with approximately 10 times lower photodetector voltages recorded for all four lasers. Polished stainless steel was found to perform better than polished aluminum in terms of reflected laser intensity. Additionally, stainless steel is less prone to surface reactions at elevated temperatures and therefore less sensitive to surface degradation. As a result, polished stainless steel was chosen as the most suitable reflector material for the single-ended sensor. To assess the effect of surface orientation on signal quality, the reflector was perturbed from its nominal orientation parallel to the sapphire window surface by ±5◦ . We observed that the signal quality did not significantly change within this range of reflector orientation, demonstrating that this sensor is insensitive to imperfect reflector alignment. This is likely due to the short path of 0.76 cm between the window and the reflector surface—the sensor is expected to become more sensitive to reflector alignment and beam steering as path length is increased. The effects of soot deposition on reflector performance were also investigated. To assess this, the stainless steel plate was placed above a sooting candle until a noticeable tarnishing of the metal surface was observed. A 15% reduction in detector voltage
Research Article
for all four lasers was observed, which was acceptable given that the reflected radiation from the untarnished stainless steel was able to fully saturate the photodetector chips in the absence of supplemental neutral-density attenuators along the collection path. We emphasize that the effects of suboptimal reflector surface quality on signal quality can be significantly mitigated by optimizing the fiber coupling efficiency or by employing lasers with sufficiently high output power. It should be noted that it is possible for some of the laser radiation to reflect from the surfaces of the sapphire window and then collected by the photodetectors without ever interacting with the combustion gases. If this stray reflection is not eliminated or accounted for, the concentration and temperature measurements will be inaccurate since the Beer-Lambert relation (Eqs. 1 and 2) assumes that all of the collected light interacts with the gas sample. In the current setup, stray reflection was eliminated from the sensor by placing the sapphire window such that the wedged surface faces away from the combustion channel. In this configuration, the incident light from the fiber bundle was reflected by the sapphire window surface at an oblique angle with respect to the collection optics and thus was not collected by the photodetectors. This was verified by operating the sensor in free space in the absence of a reflective surface and observing that none of the laser radiation was visibly detected by the photodetectors.
Applied Optics
7
(a)
4. SENSOR VALIDATION AND DEMONSTRATION Two experiments were conducted to validate and demonstrate this sensor. First, the single-ended design and the spectroscopic model used to convert the measured scanned-WMS-2 f /1 f signals into concentration values was validated by taking measurements of known quantities of CO at room temperature in a static cell fitted with a stainless steel reflector. The CO sensor was the only sensor that was validated in this work because it was the only species with sufficient absorbance and whose concentration can be well controlled at room termperature. The low E” H2 O transition is strong, but water vapor concentration is difficult to control in a static cell environment at room temperature. Nonetheless, since the WMS post-processing methods are identical for all four targeted transitions and the sensor was operated as if in an actual multi-species combustion environment, the CO validation measurements are expected to demonstrate sensor performance. Validation measurements for the two-color H2 O thermometry sensor were previously carried out by Goldenstein et al. [35] and hence were not repeated in this work. The sensor was then demonstrated in a simulated combustion exhaust environment by taking simultaneous multi-species and temperature measurements in the exhaust of a laboratory-scale CH4 -air flat-flame burner at varying equivalence ratios. A. Experimental setup
Figs. 5a and 5b show the experimental configurations used for the CO static cell validation and and flat-flame burner measurements, respectively. In the CO validation configuration, a certified standard 1.31% CO in N2 balance mixture (Praxair, Inc.) was connected to a room-temperature gas cell with 3◦ wedged BaF2 windows on opposing ends and an internal length of 2.86 cm. Light from the sensor assembly was launched through the cell, reflected off a stainless steel reflector on the opposite end, and returned to the sensor assembly for detection. The total round-trip optical path length is thus L = 5.72 cm. The gas cell was connected to a pressure transducer (MKS Instruments,
(b)
Fig. 5. Schematics of the flow facilities and optical configura-
tions for WMS measurements in (a) the CO static cell and (b) the CH4 -air flat flame burner experiment.
Inc.) and vacuum pump to allow for fine control of the pressure within the gas cell. In the burner configuration, a water-cooled flat-flame burner with a 1.52 cm wide, 15.2 cm long stainless steel mesh flameholder was placed underneath the sensor assembly at a height 1.5 cm below the center of the sapphire window with flow direction transverse to the measurement path. The burner was fed by two mass flow controllers with one carrying room air (labeled FC1) and the other carrying 99.99% CH4 (Praxair, Inc., labeled FC2). The two streams were allowed to thoroughly mix prior to entering the burner; flame equivalence ratio was controlled by adjusting the relative flow rates of each flow controller. Typical total flow rates of 30 liters per minute were achieved, leading to exhaust velocities of 0.86 m/s at a nominal burnedgas temperature of 1100 K. To reduce temperature and species non-uniformity within the measurement path, the window surface protruded approximately 0.74 cm into the edge of the flame. Additionally, the window plug and reflector were allowed to heat up in the flame prior to taking measurements. A polished
Research Article
Transition: CO P(21) P = 1.47 atm, T = 293 K
Measurement Up−scan Voigt−fit Down−scan Voigt−fit
WMS−2f/1f [A.U.]
0.25
WMS measured CO partial pressure [Torr]
0.3
Applied Optics
0.2 0.15 Turn− around
0.1 0.05
Residual
0 0.01 0.00 −0.01 0.2
0.4
0.6
0.8 Time [ms]
1
1.2
1.4
(a)
16
8
Ideal Measurement
14 12 10 8 6 4 4
6
8 10 12 14 Known CO Partial Pressure [Torr]
16
(b)
Fig. 6. (a) Measured (black) and best-fit (red and blue) WMS-2 f /1 f spectrum for the CO P(21) transition at P = 1.47 atm, L = 5.72 cm, and T = 293K. (b) Measured CO partial pressure plotted against known CO partial pressure. All measurements were taken using a 1.31% CO/N2 balance gas mixture.
stainless steel reflector was placed parallel the opposing edge of the flame-holder, leading to an optical path of L = 1.56 cm. The nominal laser frequency was maintained using temperature controllers (ILX Lightwave LDC-3900 for the lasers targeting the water and CO2 transitions, Alpes Laser Model TC-3 for the QCL targeting the CO transition), while fine control of the output frequency and intensity was achieved by modulating the injection current to the lasers with a high-compliance current controller (ILX Lightwave LDC-3900 for the lasers targeting the water and CO2 transitions, ILX Lightwave 3232 for the QCL targeting the CO transition). The lasers targeting the lower E” H2 O and CO transitions were sinusoidally modulated using f m = 90 kHz whereas the lasers targeting near the high E” H2 O the CO2 transitions were modulated using f m = 112 kHz. All four lasers were sinusoidally scanned about the linecenters of the target transitions at f s = 1 kHz, thus enabling a measurement rate of 2 kHz (two measurements per scan cycle). A National Instruments Model PXIe-1062Q data acquisition was used to generate the laser tuning signals and to record the photodetector voltages at a sampling rate of 10 MHz. Laser intensity and optical frequency tuning parameters (as defined in Eq. 11 and Eq. 12) were characterized using standard procedures [44] involving a crystalline germanium Fabry-P´erot interferometer with a free-spectral-range, ∆νFSR (cm−1 ), calculated using the following: 1 ∆νFSR = (13) 2n Ge l where l is the cavity length and nGe is the index of refraction of Germanium evaluated at the mean wavelengths of each laser using dispersion relations from Icenogle et al. [45]. A digital lock-in filter coupled with a brick-wall low-pass filter with cutoff frequency at 10 f s (i.e. 10 kHz) was applied to the raw detector signals to extract the harmonics of the WMS lineshapes for each laser. 10 f s was chosen as the cutoff frequency in order to capture a sufficient portion of the WMS-2 f /1 f frequency content within each scan cycle while also preventing frequency-domain crosstalk between the signals from each laser pair. Modulation depths (am in Eq. 11) were chosen to be 2.2 times
the half-width-half-maxima of the four transitions at the target thermodynamic conditions. This maximizes the WMS-2 f SNR as recommended in [44]. At P = 1 atm and T = 1200 K similar to the conditions measured in the flat-flame burner, modulation depths were 0.331, 0.010, 0.069, and 0.043 cm−1 for the low E” H2 O, high E” H2 O, CO2 , and CO transitions, respectively. Scan depths (as ) were chosen to encompass a significant portion of the transition lineshapes and were typically around 0.2 cm−1 . B. Measurements of CO in a static cell at room temperature
To validate the Voigt spectral-fitting WMS technique used to convert the measured WMS signals into concentration measurements, the WMS spectra of the CO P(21) transition was measured in the room temperature static cell at T = 293 K at a variety of total pressures between 0.5 and 1.5 atm. Fig. 6a shows one such set of measured (black line) and Voigt spectral-fitted (red and blue dotted lines) WMS-2 f /1 f lineshapes over a single scan period at a total pressure of 1.47 atm. The two large features on either side of the optical tuning turn-around point at 0.8 ms (indicated by the green arrow) each correspond to a CO measurement with separate Voigt profiles fitted to each feature. The free parameters in the fitting routine are the mean laser tuning frequency, the CO partial pressure, and the collisional linewidth. As can be seen, there is reasonable agreement between the measured and spectral-fitted WMS-2 f /1 f lineshapes, demonstrating that the WMS simulation model was able to accurately reproduce the measured WMS lineshapes. Additionally, despite qualitative differences between the up- and down-scan WMS-2 f /1 f features which are expected due to laser intensity variations across the scan, no systematic differences between the best-fit integrated absorbances of the up- and down-scan features were observed throughout each scan cycle. Fig. 6b compares the measured CO partial pressure, PCO , against known values calculated based on pressure transducer readings and the known CO mole fraction within the gas mixture. Each data point represents the mean CO partial pressure collected over a 10 ms interval, equivalent to 20 up- and downscans across the CO transition. The standard deviations (σ) in
Research Article
Applied Optics
0.2
Transition: Low E" H2O
WMS−2f/1f [A.U.]
0.35
T = 1126 K, P = 1 atm χ = 17.5% H2O
0.3 0.25 0.2
Turn− around
0.15
Transition: High E" H2O
Measurement Up−scan Voigt−fit Down−scan Voigt−fit
T = 1126 K, P = 1 atm χ = 17.5% WMS−2f/1f [A.U.]
0.4
Measurement Up−scan Voigt−fit Down−scan Voigt−fit
0.15
9
H2O
0.1 Turn− around 0.05
0.1
0 0.05
Residual
Residual
0.05
0 −0.05 0.2
0.4
0.6
0.8 Time [ms]
1
1.2
1.4
(a)
0 0.01 0 −0.01 0.2
0.4
0.6
0.8 Time [ms]
1
1.2
1.4
(b)
WMS−2f/1f [A.U.]
0.4
0.7
Transition: CO2 ν3 R(92)
Measurement Up−scan Voigt−fit Down−scan Voigt−fit
T = 1126 K, P = 1 atm χCO2 = 8.3%
0.6 WMS−2f/1f [A.U.]
0.5
0.3 0.2
Turn− around
Transition: CO P(21) T = 1126 K, P = 1 atm χ = 2.3%
Measurement Up−scan Voigt−fit Down−scan Voigt−fit
CO
0.5 0.4 0.3
Turn− around
0.2
0.1
(c)
0 0.02
0 Residual
Residual
0.1
0 −0.02 0.2
0.4
0.6
0.8 Time [ms]
1
1.2
1.4
0.05 0.00 −0.05 0.2
0.4
0.6
0.8 Time [ms]
1
1.2
1.4
(d)
Fig. 7. Measured (black) and best-fit (red and blue) WMS-2 f /1 f spectra for the (a) low E” H2 O transition at 2.55 µm, (b) high E” H2 O transition at 2.48 µm, (c) CO2 transition at 4.15 µm, and (d) CO transition at 4.86 µm. These measurements represent simultaneous measurements at a location 1.5 cm above the flame-holder of a laminar flat-flame CH4 -air burner at P = 1 atm. The inferred temperature is T = 1126 K based on two-color thermometry. The total path length, L, of the single-ended absorption sensor is 1.56 cm.
Research Article
0.4
χH
Applied Optics
1300 Solid lines indicate equilibrium values
O
2
χCO
2
5χCO
1125
0.2
950
0.1
775
0
0.7
0.8
0.9 1 1.1 Equivalence Ratio
1.2
Temperature [K]
Mole Fraction
0.3
600 1.3
Fig. 8. 10 ms-averaged temperature (red) and H2 O (blue), CO2 (black), and CO (green) mole fractions as a function of equivalence ratio measured 2.5 cm above the flame-holder of a laminar flat-flame CH4 -air burner at P = 1 atm. The solid lines represent the equilibrium mole fractions for each species at a temperature of 1100 K. The total path length, L, of the singleended absorption sensor is 1.56 cm.
measured PCO over the 10 ms intervals did not exceed 0.5% of the mean value for all measurements. The measured PCO agree with the known partial pressures to within 1% for all measurements except for the PCO = 13.9 torr (total P = 1.4 atm) measurement, which had a discrepancy of 3%. This discrepancy can be explained by the reduced SNR at elevated pressures. Since WMS-2 f /1 f is most sensitive to the curvature of the absorption spectra at transition linecenter, the increased broadening and therefore lower absorption curvature at higher pressures results in lower measurement accuracy. In general, however, the PCO measurements agree well with the known values, thus validating the single-ended sensing strategy and spectral-fitting technique employed in this work. C. Measurements in the exhaust of a CH4 -air flat flame burner
An initial demonstration of the single-ended sensor in a combustion environment was carried out by performing simultaneous multi-species and temperature measurements in the exhaust of the CH4 -air flat-flame burner at atmospheric pressure. Measurements were taken at global equivalence ratios between 0.7 and 1.2. Prior to the measurements, the detection setup was purged with dry N2 until the absorption due to the low E” H2 O feature visually reached steady state. The residual absorption due to the remaining humidity within the optical path was removed from the measurement by the background subtraction technique as utilized by several previous researchers [5, 30, 46, 47]. In this technique, instead of using the ideal non-absorbing laser intensity tuning characteristics as expressed by Eq. 12 to simulate the WMS-2 f /1 f spectra, the intensity tuning waveform that included the absorption due to residual humidity was used as recommended in [29]. This waveform was generally measured after the humidity purge has reached steady state but prior to the flame measurements. Implicit in this technique is the assumption that ambient humidity remains constant throughout the measurements, which is reasonable given the short amount of time (∼ 10 min) needed to complete the measurements.
10
Fig. 7 shows a set of measured (black line) and Voigt-fitted (red- and blue-dashed lines) WMS-2 f /1 f spectra for each of the four transitions at an equivalence ratio of 1.2 for one scan cycle. Here, the features on either side of the wavelength turn-around point at approximately 8 ms (indicated by the green arrows) each correspond to a separate measurement of the target species. As can be seen, the Voigt-fitted WMS-2 f /1 f spectra match the measured WMS-2 f /1 f spectra well for all four transitions. Note that an accurate collisional broadening model is not needed since the Voigt collisional linewidths are fitted parameters in the WMS spectral-fitting routine. As discussed in Section C, the converged values for integrated absorbance for each WMS feature was used to calculate the temperature and mole fraction using Eqs. 8 and 10. For the particular measurements shown in Fig. 7, the 10 ms-averaged temperature was found to be 1126 K with H2 O, CO2 , and CO mole fractions of 17.5%, 8.3%, and 2.3%, respectively. Fig. 8 shows the measured temperatures and H2 O, CO2 , and CO mole fractions (shown as discrete points) across the range of equivalence ratios measured in this experiment. Each data point corresponds to the averaged value measured over 10 ms, or 20 up- and down-scans of each transition. For comparison, equilibrium mole fractions (calculated using Cantera [48]) for each species was evaluated at the average measured temperature of 1100 K and are displayed as solid lines. Adiabatic flame temperatures were not compared against the measured temperatures since the flame was highly non-adiabatic. Measured H2 O and CO2 mole fractions agreed well with equilibrium calculations in general, whereas CO measurements had significant discrepancies near φ = 1. Two major factors contribute to this discrepancy. First, low measurement SNR due to low CO concentrations at near-stoichiometric operating conditions are expected to significantly affect the results at φ = 1.05 and 1.1. Second, CO concentrations in the exhaust of flames near stoichiometric conditions are highly sensitive to temperature histories [49], which can lead to significant differences between equilibrium and actual CO concentrations. D. Detection limits
We now consider the detection limits of this sensor. This requires a quantification of SNR in the WMS-2 f /1 f signals, which for this work was defined as the square root of the sum of the spectral power of the WMS-2 f harmonic features divided by the sum of the spectral power of the non-harmonic noise features within the 2 f spectra. Mathematically, this can be expressed as follows: v u 10 2 2 u ˆ ˆ u ∑ S ( ξ = m f ) S ( ξ = m f ) − 2 f ,bg s s u m=−10 2 f SNR = u 2 u t ∑ Sˆ2 f (ξ 6= m f s ) ξ
(14) Here, Sˆ2 f and Sˆ2 f ,bg are the discrete Fourier transforms of the measurement and non-absorbing background signals, respectively, with the lock-in filter applied at 2nd harmonic (2 f ) of the modulation frequency. ξ is the temporal frequency variable and m is some integer such that m f s is within the cutoff frequency bounds of the lock-in filter (for this sensor, −10 ≤ m ≤ 10 since the cutoff frequency was set at 10 f s ). In this definition, the numerator represents the signal power due to laser modulation and absorption whereas the denominator represents the spectral power of the noise floor near the 2nd harmonic. Note that contributions to the 2 f signal power due to laser tuning non-
Research Article
Applied Optics
−5
10
Spectral Power Lock−in filter cutoff bounds
Spectral features due to absorption
- -2 -ˆ -S2f -
−10
10
−15
10
Spectral features due to noise
−20
−10
0 ξ [kHz]
10
20
Fig. 9. Spectral power at the second harmonic of the low E”
H2 O transition for the φ = 1.2 CH4 -air flat-flame burner measurement. Lock-in filter cutoff frequencies are shown in reddotted lines while spectral features due to absorption and noise are labelled with green and blue arrows, respectively. linearities and other non-absorption effects on laser intensity were removed from the SNR definition by subtracting the background term in the numerator of Eq. 14, thereby leaving only the absorption-induced contributions to signal power. Fig. 9 provides a visualization of this definition of SNR. Here, 2 the spectral power, Sˆ2 f (black solid line), of the WMS-2 f signal due to the low E” H2 O transition is shown for the φ = 1.2 measurement in Fig. 7. The red-dotted lines represent the cutoff frequencies of the lock-in filter; only the frequency content within these cutoff bounds was considered when evaluating Eq. 14. A few of the spectral features due to absorption and noise are labeled with green and blue arrows, respectively. Note that the measurement duration of 10 ms is an integer multiple of the scan period, hence we need not be worried about spectral leakage from the WMS signal features into the noise bands when applying the discrete Fourier transform. For the φ = 1.2 measurement, SNR was calculated to be 127 for the low E” H2 O transition, 62 for the high E” H2 O transition, 72 for the CO2 transition, and 31 for the CO transition. Assuming that an SNR of 5 is required for meaningful measurements and that SNR is approximately proportional to peak absorbance (this is generally the case in optically thin environments), we estimate that the detection limits in terms of mole fraction are 1.4% for H2 O, 0.6% for CO2 , and 0.4% for CO.
5. CONCLUSIONS The development, validation, and initial demonstration of an in-situ single-ended back-reflection-based LAS sensor was presented for the simultaneous measurement of H2 O, CO2 , and CO concentration and temperature in combustion flows. The use of the noise-tolerant scanned-WMS-2 f /1 f technique along with light sources targeting mid-IR transitions within the fundamental rovibrational bands of the combustion species allowed the sensor to perform high SNR measurements of the gas properties over short path lengths on the order of 1 cm. The sensor was demonstrated in a laboratory-scale CH4 -air flat-flame burner at varying equivalence ratios, achieving theoretical detection
11
limits of 1.4% H2 O, 0.6% CO2 , and 0.4% CO at a measurement rate of 2 kHz. Measured species mole fractions agreed well with theoretical values calculated using a chemical equilibrium solver. This work represents the first demonstration of a multi-species single-ended LAS sensor operating in the mid-IR. Broadly, we envision this sensor to extend the applicability of LAS sensors to previously optically-inaccessible combustion environments such as within the turbomachinery of jet engines. Work is currently underway to increase the measurement rate of the current sensor to high-speed combustion processes such as detonation waves in RDEs as demonstrated by Rein et al. [25]. Additionally, this type of sensor architecture can be extended to measure velocity and additional species (e.g. trace pollutants such as NOx ), which can provide important diagnostics information such as enthalpy flow rate and emissions quantification for many practical combustion systems. Funding. Small Business Innovation Research/Small Business Technology Transfer (SBIR/STTR) (N15A-T021); Air Force Office of Scientific Research (AFOSR) (FA9550-15-1-0049). Acknowledgement. The authors would like to thank D. Dausen of the Naval Postgraduate School (NPS) for contributions to sensor design. The authors would also like to thank Dr. John Hoke of Integrated Science Solutions, Inc. and Dr. Chiping Li of AFOSR for monitoring the SBIR/STTR and AFOSR contracts, respectively.
REFERENCES 1. R. K. Hanson, “Applications of quantitative laser sensors to kinetics, propulsion and practical energy systems,” Proceedings of the Combustion Institute 33, 1–40 (2011). 2. S. T. Sanders, J. A. Baldwin, T. P. Jenkins, D. S. Baer, and R. K. Hanson, “Diode-laser sensor for monitoring multiple combustion parameters in pulse detonation engines,” Proceedings of the Combustion Institute 28, 587–594 (2000). 3. J. T. C. Liu, G. B. Rieker, J. B. Jeffries, M. R. Gruber, C. D. Carter, T. Mathur, and R. K. Hanson, “Near-infrared diode laser absorption diagnostic for temperature and water vapor in a scramjet combustor,” Applied Optics 44, 6701 – 6711 (2005). 4. G. B. Rieker, H. Li, X. Liu, J. T. C. Liu, J. B. Jeffries, R. K. Hanson, M. G. Allen, S. D. Wehe, P. A. Mulhall, H. S. Kindle, A. Kakuho, K. R. Sholes, T. Matsuura, and S. Takatani, “Rapid measurements of temperature and H2O concentration in IC engines with a spark plug-mounted diode laser sensor,” Proceedings of the Combustion Institute 31 II, 3041–3049 (2007). 5. X. Chao, J. B. Jeffries, and R. K. Hanson, “Development of laser absorption techniques for real-time, in-situ dual-species monitoring (NO/NH3, CO/O2) in combustion exhaust,” Proceedings of the Combustion Institute 34, 3583–3592 (2013). 6. K. Sun, R. Sur, X. Chao, J. B. Jeffries, R. K. Hanson, R. J. Pummill, and K. J. Whitty, “TDL absorption sensors for gas temperature and concentrations in a high-pressure entrained-flow coal gasifier,” Proceedings of the Combustion Institute 34, 3593–3601 (2013). 7. C. S. Goldenstein, “Wavelength-modulation spectroscopy for determination of gas properties in hostile environments,” Thesis (2014). 8. R. M. Spearrin, C. S. Goldenstein, I. A. Schultz, J. B. Jeffries, and R. K. Hanson, “Simultaneous sensing of temperature, CO, and CO2 in a scramjet combustor using quantum cascade laser absorption spectroscopy,” Applied Physics B 117, 689–698 (2014). 9. R. Sur, K. Sun, J. B. Jeffries, J. G. Socha, and R. K. Hanson, “Scannedwavelength-modulation-spectroscopy sensor for CO, CO2, CH4 and H2O in a high-pressure engineering-scale transport-reactor coal gasifier,” Fuel 150, 102–111 (2015). 10. M. E. Webber, D. S. Baer, and R. K. Hanson, “Ammonia monitoring
Research Article
near 1.5 µm with diode-laser absorption sensors,” Applied optics 40, 2031–2042 (2001). 11. C. L. Strand and R. K. Hanson, “Quantification of supersonic impulse flow conditions via high-bandwidth wavelength modulation absorption spectroscopy,” AIAA Journal 53, 2978–2987 (2015). 12. W. Y. Peng, R. Sur, C. L. Strand, R. M. Spearrin, J. B. Jeffries, and R. K. Hanson, “High-sensitivity in-situ QCLAS-based ammonia concentration sensor for high-temperature applications,” Applied Physics B: Lasers and Optics 122, 1–12 (2016). 13. L. C. Philippe and R. K. Hanson, “Laser diode wavelength-modulation spectroscopy for simultaneous measurement of temperature, pressure, and velocity in shock-heated oxygen flows.” Applied optics 32, 6090– 6103 (1993). 14. J. Wolfrum, “Lasers in combustion: From basic theory to practical devices,” Symposium (International) on Combustion 27, 1–41 (1998). 15. S. D. Wehe, “Development of a tunable diode laser probe for measurements in hypervelocity flows,” Ph.D. thesis (2000). 16. G. Wysocki, A. A. Kosterev, and F. K. Tittel, “Spectroscopic trace-gas sensor with rapidly scanned wavelengths of a pulsed quantum cascade laser for in situ NO monitoring of industrial exhaust systems,” Applied Physics B: Lasers and Optics 80, 617–625 (2005). 17. V. Ebert, H. Teichert, P. Strauch, T. Kolb, H. Seifert, and J. Wolfrum, “Sensitive in situ detection of CO and O2 in a rotary kiln-based hazardous waste incinerator using 760 nm and new 2.3 µm diode lasers,” Proceedings of the Combustion Institute 30, 1611–1618 (2005). 18. C. H. Smith, C. S. Goldenstein, and R. K. Hanson, “A scannedwavelength-modulation absorption-spectroscopy sensor for temperature and H2O in low-pressure flames,” Measurement Science and Technology 25, 1–13 (2014). 19. Z. Wang and S. T. Sanders, “Toward single-ended absorption spectroscopy probes based on backscattering from rough surfaces: H2O vapor measurements near 1350 nm,” Applied Physics B 121, 187–192 (2015). 20. Z. Wang, S. T. Sanders, and M. A. Robinson, “Spatially resolved concentration measurements based on backscatter absorption spectroscopy,” Applied Physics B 122, 1–6 (2016). 21. M. B. Frish, R. T. Wainner, J. Stafford-Evans, B. D. Green, M. G. Allen, S. Chancey, J. Rutherford, G. Midgley, and P. Wehnert, “Standoff sensing of natural gas leaks: Evolution of the remote methane leak detector (RMLD),” (2004), pp. 1941–1943. 22. M. B. Frish, R. T. Wainner, M. C. Laderer, B. D. Green, and M. G. Allen, “Standoff and miniature chemical vapor detectors based on tunable diode laser absorption spectroscopy,” IEEE Sensors Journal 10, 639– 646 (2010). 23. J. Chen, A. Hangauer, R. Strzoda, and M. C. Amann, “Laser spectroscopic oxygen sensor using diffuse reflector based optical cell and advanced signal processing,” Applied Physics B: Lasers and Optics 100, 417–425 (2010). 24. C. S. Goldenstein, R. M. Spearrin, and R. K. Hanson, “Fiber-coupled diode-laser sensors for calibration-free stand-off measurements of gas temperature, pressure, and composition,” Applied Optics 55, 2–7 (2015). 25. K. D. Rein and S. Roy, “Time-resolved in-situ absorption spectroscopy of a hydrogenair rotating detonation engine using a fiber-coupled tunable laser system,” in “AIAA SciTech,” (2016), January, pp. 1–10. 26. U. Platt and J. Stutz, Differential optical absorption spectroscopy (Springer Berlin Heidelberg, 2008). 27. P. Kluczynski and O. Axner, “Theoretical description based on Fourier analysis of wavelength-modulation spectrometry in terms of analytical and background signals.” Applied optics 38, 5803–5815 (1999). 28. R. K. Hanson, R. M. Spearrin, and C. S. Goldenstein, Spectroscopy and optical diagnostics for gases (Springer, New York, 2015), 1st ed. 29. C. S. Goldenstein, C. L. Strand, I. A. Schultz, K. Sun, J. B. Jeffries, and R. K. Hanson, “Fitting of calibration-free scanned-wavelength-modulation spectroscopy spectra for determination of gas properties and absorption lineshapes.” Applied optics 53, 356–367 (2014). 30. K. Sun, X. Chao, R. Sur, C. S. Goldenstein, J. B. Jeffries, and R. K. Hanson, “Analysis of calibration-free wavelength-scanned wavelength modulation spectroscopy for practical gas sensing using tunable diode
Applied Optics
12
lasers,” Measurement Science and Technology 24, 1–12 (2013). 31. C. L. Strand, “Scanned wavelength modulation absorption spectroscopy with application to hypersonic impulse flow facilities,” Ph.D. thesis (2014). 32. K. Levenberg, “A method for the solution of certain non-linear problems in least squares,” Quarterly of Applied Mathematics 2, 164–168 (1944). 33. D. W. Marquardt, “An algorithm for least-squares estimation of nonlinear parameters,” Journal of the Society for Industrial and Applied Mathematics 11, 431–441 (1963). 34. D. F. Dausen, C. M. Brophy, R. G. Wright, and J. D. Marder, “Design of an optically-accessible rotating detonation engine,” in “48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit,” (2012), pp. 1–12. 35. C. S. Goldenstein, I. A. Schultz, R. M. Spearrin, J. B. Jeffries, and R. K. Hanson, “Scanned-wavelength-modulation spectroscopy near 2.5 µm for H2O and temperature in a hydrocarbon-fueled scramjet combustor,” Applied Physics B 116, 717–727 (2014). 36. C. S. Goldenstein, C. A. Almodóvar, J. B. Jeffries, R. K. Hanson, and C. M. Brophy, “High-bandwidth scanned-wavelength-modulation spectroscopy sensors for temperature and H2O in a rotating detonation engine,” Measurement Science and Technology 25, 1–11 (2014). 37. C. S. Goldenstein, R. M. Spearrin, J. B. Jeffries, and R. K. Hanson, “Wavelength-modulation spectroscopy near 2.5 µm for H2O and temperature in high-pressure and -temperature gases,” Applied Physics B: Lasers and Optics 116, 705–716 (2014). 38. R. M. Spearrin, C. S. Goldenstein, J. B. Jeffries, and R. K. Hanson, “Quantum cascade laser absorption sensor for carbon monoxide in highpressure gases using wavelength modulation spectroscopy.” Applied optics 53, 1938–46 (2014). 39. C. S. Goldenstein, J. B. Jeffries, and R. K. Hanson, “Diode laser measurements of linestrength and temperature-dependent lineshape parameters of H2O-, CO2-, and N2-perturbed H2O transitions near 2474 and 2482nm,” Journal of Quantitative Spectroscopy and Radiative Transfer 130, 100–111 (2013). 40. R. M. Spearrin, W. Ren, J. B. Jeffries, and R. K. Hanson, “Multi-band infrared CO2 absorption sensor for sensitive temperature and species measurements in high-temperature gases,” Applied Physics B: Lasers and Optics 116, 855–865 (2014). 41. L. Rosenmann, S. Langlois, and J. Taine, “Diode laser measurements of CO2 hot nand line intensities at high temperature near 4.3 µm,” Journal of Molecular Spectroscopy 158, 263–269 (1993). 42. P. Varghese and R. Hanson, “Tunable infrared diode laser measurements of line strengths and collision widths of 12C16O at room temperature,” Journal of Quantitative Spectroscopy and Radiative Transfer 24, 479–489 (1980). 43. L. Rothman, I. Gordon, R. Barber, H. Dothe, R. Gamache, A. Goldman, V. Perevalov, S. Tashkun, and J. Tennyson, “HITEMP, the hightemperature molecular spectroscopic database,” Journal of Quantitative Spectroscopy and Radiative Transfer 111, 2139–2150 (2010). 44. G. B. Rieker, J. B. Jeffries, and R. K. Hanson, “Calibration-free wavelength-modulation spectroscopy for measurements of gas temperature and concentration in harsh environments.” Applied optics 48, 5546–5560 (2009). 45. H. W. Icenogle, B. C. Platt, and W. L. Wolfe, “Refractive indexes and temperature coefficients of germanium and silicon,” Applied optics 15, 2348–2351 (1976). 46. R. Sur, K. Sun, J. B. Jeffries, and R. K. Hanson, “Multi-species laser absorption sensors for in situ monitoring of syngas composition,” Applied Physics B: Lasers and Optics 115, 9–24 (2014). 47. K. Owen and A. Farooq, “A calibration-free ammonia breath sensor using a quantum cascade laser with WMS 2f/1f,” Applied Physics B 116, 371–383 (2013). 48. D. G. Goodwin, H. K. Moffat, and R. L. Speth, “Cantera: An objectoriented software toolkit for chemical kinetics, thermodynamics, and transport processes,” (2016). 49. M. Schoenung and R. K. Hanson, “CO and temperature measurements in a flat flame by laser absorption spectroscopy and probe techniques,” Combustion Science and Technology 24, 227–237 (1980).
