air-ratio measurements with LIBS

Combustion and Flame 198 (2018) 120–129

Contents lists available at ScienceDirect

Combustion and Flame journal homepage: www.elsevier.com/locate/combustflame

High-pressure 1D fuel/air-ratio measurements with LIBS Yue Wu a, Mark Gragston a, Zhili Zhang a,∗, Paul S. Hsu b, Naibo Jiang b, Anil K. Patnaik b, Sukesh Roy b, James R. Gord c a

Department of Mechanical, Aerospace, and Biomedical Engineering, University of Tennessee, Knoxville, TN 37996, USA Spectral Energies LLC, 4065 Executive Dr., Beavercreek, OH 45430, USA c Air Force Research Laboratory, Aerospace Systems Directorate, Wright-Patterson AFB, OH 45433, USA b

a r t i c l e

i n f o

Article history: Received 3 May 2018 Revised 23 June 2018 Accepted 6 September 2018

Keywords: Turbulent flame Fuel–air ratio One-dimensional Nanosecond Laser-induced breakdown spectroscopy High pressure

a b s t r a c t Quantitative, one-dimensional (1D), single-laser-shot, fuel–air ratio (FAR) measurements in both laminar and turbulent methane–air flames were conducted using time-gated nanosecond-laser-induced breakdown spectroscopy (ns-LIBS) line imaging. In the laminar methane–air flames at a pressure of 1–11 bar, hydrogen (Hα ) and nitrogen (NII ) atomic emission lines at 568 and 656 nm, respectively, were selected to establish a correlation between the line intensities and the local FAR. The spatial calibration profiles of the N/H ratios in the flames at various pressures were obtained in one dimension. The effects of the laser energy and pressure on the stability and precision of the 1D FAR measurements were investigated. It was observed that the N/H correlation is significantly reduced at ∼11 bar, which sets the limits of the 1D LIBSbased FAR measurements. Single-laser-shot 1D FAR measurements were conducted in a turbulent flame at atmospheric pressure, and multiline LIBS was performed to extend the measurement area of interest. Spatially and spectrally resolved line LIBS can provide the local FAR with a spatial resolution of ∼0.1 mm. These results hold promise for the utilization of ns-LIBS for spatially resolved 1D FAR measurements in turbulent flames at elevated pressures. © 2018 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

1. Introduction Nanosecond-laser-induced breakdown spectroscopy (ns-LIBS) has been widely applied in scientific and engineering fields, ranging from basic science laboratories to outer space [1]. The optical breakdown due to a focused high-intensity laser beam in solids, liquids, and gases reveals the elemental information of the target with minimal or no requirements for sample preparation. The breakdown process is initiated with the creation of seed electrons near the laser focal position by the front end of the high-intensity pulse [2,3]. The seed electrons, which are generated by a multiphoton ionization process, absorbs the remainder of the same laser pulse through an inverse Bremsstrahlung process and causes avalanche ionization with a rapid growth of the electron density. Hence, the laser-induced breakdown (LIB) process is considered as a probabilistic process [4,5]. In recent decades, the LIBS technique has been utilized for combustion diagnostics based on direct measurements of individual emission lines, delivering a significant amount of quantitative information including the elemental composition, species con-

∗

Corresponding author. E-mail address: [email protected] (Z. Zhang).

centration [6–11], flame density [7,10], temperature [12,13], and fuel–air ratio (FAR) [14–25]. The atomic emissions from a laserinduced plasma are utilized to infer the quantitative composition and species concentrations in the flames [18]. In general, the dependence of the emissions of the laser-induced gas breakdown on the laser energy (LE), the composition of the target gas, and the ambient environment causes significant fluctuations in the spectral line intensity. These factors hinder the reliability of the emission signal for direct correlation to the actual mole fraction or species concentration [26,27]. Do et al. [7] utilized LIBS for the measurement of the gas density based on the direct correlation between the LE scattered/absorbed by a plasma and the density. However, there are significant challenges regarding the measurement accuracy with high laser scattering/absorption by the plasma in real combustion environments due to laser-beam steering, collisional quenching, and the generation of soot particles. Lee and Hedge [12] derived the temperature from LIBS spectra, assuming that the nitrogen signal intensity is proportional to the density or inversely proportional to the temperature. Kiefer et al. [13] took advantage of the breakdown threshold being a strong function of the density, which provides a means for thermometry. Typically, large shot-to-shot signal (i.e., spectral intensity) fluctuations are expected for LIBS-based combustion diagnostics using longer laser pulses. The high signal fluctuations originating from the stochas-

https://doi.org/10.1016/j.combustflame.2018.09.009 0010-2180/© 2018 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Y. Wu et al. / Combustion and Flame 198 (2018) 120–129

tic behavior of the gas breakdown process hinder the development of LIBS applications for the quantitative measurement of the species concentration, density, and temperature in combustion environments. Several studies have demonstrated that the ratios of the peaks of emission lines can be used to correlate and quantify the FAR (or equivalence ratio) in premixed combustion systems [6,14,16– 18,28,29] and practical engines [15,19,20,30–32]. Their results also demonstrated that by taking the ratio of signals from various emission lines, the measurement fluctuations can be significantly reduced [14,28,33]. Various emission lines have been used for FAR measurements, including Hα (656 nm)/NII (568 nm) [6,34], Hα (656 nm)/NI (500 nm) [11,24,33], Hß (486 nm)/NI (746 nm) [31], Hα (656 nm)/OI (777 nm) [8,14,16,17,20,22–25,28,29,32,33,35,36], Hα (656 nm)/NI (742–746 nm) [20,22,23], CI (711 nm)/NI (740– 750 nm) [15], CI (711 nm)/OI (777 nm) [15], CI (711 nm)/(NI (746 nm) + OI (777 nm)) [18], CN (707–734 nm)/NI (740–750 nm) [15], and CN (707–734 nm)/OI (777 nm) [15]. By weighing the advantages and disadvantages for different applications and scenarios, most of the experiments employed time-gated detection to suppress sampling of the plasma continuum emission and flame chemiluminescence, thereby improving the signal-to-background ratio [6,14–20,28–32]. Moreover, other studies have shown that ungated detection with appropriate blocking of the scattered beam can be used for measurements of the FAR [13,25,33]. Although the ratios of various spectral lines provide FAR measurements, all previous occurrences of the detection of the LIBSbased FAR are single-point measurements; that is, the light emitted from the plasma was conventionally treated as a single point source in those experiments. In fact, the plasma and subsequent emission are far from spherical in shape [4,37,38]. Generally, the nanosecond-laser-induced plasma volume has an ellipsoidal shape with a width of tens to hundreds of micrometers and a length of a few millimeters. Hence, the relative spatial position of the plasma recorded by the spectrally resolved detection system (i.e., a spectrometer equipped with a camera) would directly influence the experimental FAR measurements. Moreover, line LIBS has potential applications in fluid shear layers with a very fine structure (i.e., a thickness of approximately a micrometer), where both the velocity and shear stress dramatically vary near the boundary. In this study, spatially and spectrally resolved ns-LIBS was explored for FAR measurements in methane–air flames at an elevated pressure. The primary focus of this study is to expand the dimensionality of the measurement from a point to a line and to present a spatial analysis of the LIBS spectral line emissions for one-dimensional (1D) FAR measurements in turbulent flames. The three major conclusions obtained from the results presented in this paper are as follows. (1) A laminar Hencken flame in a high-pressure combustion chamber is used to comprehensively study the spatial and spectral characteristics of LIBS emission lines with respect to the LE and pressure conditions. (2) Singlelaser-shot 1D FAR measurements in a turbulent methane diffusion flame using ns-line-LIBS are demonstrated. (3) We extend singleline LIBS to multiline LIBS using a homemade linear lens array to demonstrate multiline LIBS for FAR measurements in reacting flows. This paper is organized as follows. The experimental arrangements for high-pressure FAR measurements are described in Section 2. Section 3 presents the spectral and spatial analyses of atomic line emissions in laminar Hencken flames under various pressure conditions. Section 4 presents quantitative 1D FAR measurements for atmospheric-pressure turbulent flames. Section 5 presents a feasibility study of multiline LIBS for large-area FAR measurements. The paper concludes with a summary and discussion for further applications/improvements in Section 6.

121

Fig. 1. Experimental setup for ns-LIBS in laminar flames at an elevated pressure. The various components are listed on the left side of the figure. The inset is a snapshot of a laminar methane–air flame at 6 bar.

2. Experimental setup Figure 1 shows the experimental setup for ns-LIBS measurement in a stainless-steel high-pressure combustion chamber. The dimensions of the inner chamber are 5.85 in. in diameter and 25 in. in height. The chamber has four 2-in. diameter windows for optical diagnostics. The high-pressure chamber can accommodate a 2-in.-diameter burner for the required LIBS-based 1D FAR calibration tests. The exhaust for the gases was operated via a nozzle on the top of the chamber. Air was used as a buffer gas to help push the exhaust away from the flame while also maintaining the overall pressure in the chamber. The flow rates of methane, nitrogen, and air were all controlled by digital flow meters. A cold and dry buffer gas (air or nitrogen) was used to purge the windows continuously during measurement to prevent the condensation of water on the chamber windows. A laminar flame is generated by a well-calibrated Hencken burner, which is ideal for FAR calibration experiments. A 5.08cm-diameter Hencken burner (flame area: 2.54 cm. × 2.54 cm) was installed in the high-pressure chamber, as depicted in Fig. 1. The Hencken flame is a diffusion-limited, flow-stabilized, weakly stretched premixed flame. At elevated pressures, the flame is more stretched and flame reaction zone is thinner. When the Hencken burner was operated in the chamber, the gas was exhausted from the top of the chamber through a tube to a fume hood. The chamber temperature and pressure were monitored to ensure that these conditions are maintained during measurement. A Hencken methane–air flame at 6 bar (6 × 105 Pascal) is shown in the inset in Fig. 1. For LIBS experiments, we employed the second harmonics of a Nd:YAG laser (Powerlite DLS, Continuum, Inc.) with 10-nsduration laser pulses and a maximum energy of 200 mJࢧpulse at a 10-Hz repetition rate. The shot-to-shot laser pulse fluctuation is ∼3 %. The laser pulse energy was adjusted by a half-wave plate and polarizer. The incident laser beam was focused 20 mm above the burner surface downstream of the flame front (at the center of the vessel) using a spherical lens with a focal length of + 50 mm. The probe areas were located entirely in the post combustion region for all pressure conditions. The size of the plasma is much smaller than the overall scale of the composite flame (i.e, reacting and post-reacting) structure. The length of laser-induced plasma is limited to a few millimeters while the composite flame width is ∼20 mm at the point of measurement. The LIBS emissions from the probe volume were first collected and collimated inside the vessel by a visible-light spherical achromatic lens with focal length

122


Fig. 2. Line-LIBS spectral images obtained in methane–air laminar flames with an equivalence ratio φ = 1 at 1 bar for five different input laser energies (50, 75, 100, 125, and 150 mJ/pulse) covering a 200-ns time duration of plasma evolution (40–240 ns) after the excitation laser pulse. Each individual image is both spectrally (horizontal axis) and spatially (vertical axis) resolved. Detailed information is provided in Fig. 3.

F = + 100 mm. The collimated signal beam was collected outside the vessel by using another identical achromatic lens that was then coupled onto the entrance slit of a 0.25-m spectrometer equipped with a 150-groove/mm grating blazed at 500 nm (SpectraPro 2300i, Princeton Instruments). A time-gated intensified charge-coupleddevice (ICCD) camera (PI-MAX4, Princeton Instruments) with a 1024 × 512 pixel array was used to record the spectra covering a wavelength region of 560–700 nm. The orientation of the entrance slit of the spectrometer was along the optical axis of the laser beam. Hence, the horizontal and vertical dimensions of the ICCD array provided spectral and spatial information, respectively. The laser system and ICCD camera were both externally triggered by a digital delay generator (Stanford Research Systems, DG645) to control the time elapsed between a laser pulse and signal collection (i.e., the gate delay Tdelay ). The acquisition time (i.e., the ICCD gate width) was 20 ns in the experiment. The short camera exposure time was chosen to suppress/attenuate the flame chemiluminescence and benefit from the enhanced signal-to-background ratio perspective. The gate delay (> 40 ns) allows the strong Rayleigh scattering and the scattering from metallic surfaces inside the vessel to be avoided. 3. Spectral and spatial analyses of ns-LIBS in laminar flames at an elevated pressure Unlike the random and uncontrollable spark formation in spark-induced breakdown spectroscopy (SIBS), LIB generates a laser-guided plasma line that further expands along both the welldefined laser path in the longitudinal direction and in the transverse direction around the beam waist of the laser. Hence, the laser-induced plasma has a typical profile with a width of tens or hundreds of micrometers and a plasma length of a few millimeters. In our previous studies, the stability, sensitivity, and precision of pointwise ns-LIBS-based FAR measurements have been analyzed in methane–air flames at elevated pressures (1–11 bar) [34]. In this work, we extended our analyses to confirm the validity of expanding the pointwise LIBS measurements into spatially resolved, 1D, ns-LIBS-based, FAR measurements. The optimized experimental parameters from Ref. [34] are adopted in the present work, including

the camera exposure gate delay (Tdelay = 80 ns), where the experimental setup was optimized for single-point LIBS measurements at elevated pressures. The data shown in all the figures below have taken into account of the quantum efficiency of the camera, the imaging efficiencies of grating and optical transmission profiles of achromatic optics for both H and N emissions. The residual effects from the broadband emission of the plasma were minimized by the delay, which was also been subtracted. 3.1. 1D Analyses of individual ns-LIBS line emissions in a Hencken flame at 1–11 bar The temporally resolved emissions from both plasma generation and evolution are obtained from flat Hencken flames for various gate delays and input LEs. By optimizing these two factors, the stability and precision of line-LIBS measurements at elevated pressures can be improved. Figure 2 shows a matrix of images obtained for different LEs at different camera gate delays, which shows the temporal evolution of the spectral images of line-LIBS emissions in a methane–air flame at atmospheric pressure for the first 240 ns after laser beam arrival. The image matrix covers five different input energies of the drive laser, showing the evolution of plasma emissions from a time delay of 40 ns to 240 ns after the arrival of the laser pulse. The images were averaged for 100 shots to improve the signal-to-noise ratios (SNRs). In each individual subplot, the horizontal axis spans from 560 nm to 680 nm, and the vertical axis gives the spatially resolved distribution of the line emissions (∼4 mm), as shown in the magnified plot in Fig. 3. The continuum emission from hot electrons rapidly decays as the plasma cools down. At Tdelay = 40 ns, discrete emission lines, NII (568 nm) and Hα (656 nm), are both clearly observed and identified above the broadband residue. In general, for ns-LIBS, strong broadband spectral emission initially occurs in the spectrum (for Tdelay ≤ 20 ns), and this emission exponentially decays within a short lifetime. Detailed information about these elemental emission lines is available in the NIST database [39]. In Fig. 2, the LIBS signal sustains the longest period for LEs of ∼150 mJ/pulse, which can be understood as follows. At a higher LE, a denser plasma is created, thereby increasing the lifetime of


Fig. 3. LIBS spectra obtained in methane–air laminar flames with φ = 1 at 1 bar and Tdelay = 80 ns with incident laser energies of 50 and 150 mJ/pulse.

the LIBS signal because of longer recombination processes. Both the NII (568 nm) and Hα (656 nm) line emissions show inhomogeneous line-shape distributions in the spatial domain inherited from the laser-induced plasma initiation and evolution, which will be quantified in Section 4. The Hα (656 nm) and NII (568 nm) lines respectively represent the fuel and air concentrations in the flame, which were used to measure the FAR. Figure 3 shows the spatially resolved ns-LIBS emission lines for incident laser pulses with two different energies (50 and

123

150 mJ/pulse). The y axis represents the spatial scale of the plasma emission (calibrated by the spectrometer slit width), and the x axis represents the spectral dispersion of the emissions from the spectrometer. These two images were acquired after a time delay (Tdelay ) of 80 ns with a 20-ns gate width. Both images were averaged over 100 laser shots. It was observed that both the spatial and spectral distributions of the LIBS signals vary with the incident LE because the laser fluences could affect the temperature and size of the plasma [18]. Figure 4 shows the ratio of the NII (568 nm) emission line to the Hα (656 nm) emission line at different spatial positions within the plasma region. The laser beam is focused and passing from the left-hand side (i.e., a position of − 1.2 mm) into the focal zone. The presented data were acquired at a delay of 80 ns after the arrival of the laser pulse with a 20-ns gate width and averaged over 100 laser shots. The delay and camera gate width were the same as previous single-point measurements [34]. The ratios in Fig. 4 were directly obtained under the same experimental conditions and thus reflected the fuel/air ratios of the Hencken flames. The spatial profiles of the NII /Hα ratio are fixed at each pressure, which changes only with the change in the LE. The spectrally resolved 1D LIBS measurements could be used to extend the FAR measurement capability from a point to a line, provided the spatial profiles of the local plasma are calibrated. It should be noted that the maximum peak heights for atomic emission lines were used and averaged over 5 pixels around the peak. Integrated peak area was not used to eliminate the impacts of residual broadband emissions and flame chemiluminescence. The characteristic properties of spatially and spectrally resolved 1D ns-LIBS in laminar flames at an elevated pressure are briefly summarized as follows: (1) The distribution of the NII /Hα ratio in the constant FAR Hencken flames is not identical within the LIBS plasma along the laser propagation direction. The ratios with the spatial distributions are dependent on both the pressure and incident LE, which is primarily caused by differences in the LIBS

Fig. 4. Spatially resolved NII (568 nm)/Hα (656 nm) ratios with incident laser energies of 50 and 150 mJ/pulse under different pressure conditions, showing the spatial profiles of NII /Hα in the Hencken flames at different pressures (corresponding to Fig. 3).

124


Fig. 5. (Top) Spatial distributions of the mean spectral emission intensities of Hα , NII , and the NII /Hα ratio in methane–air flames at a pressure of 1 bar. (Bottom) The corresponding relative standard deviations (RSDs) are presented. The RSD is defined as the ratio of the standard deviation to the mean value of the replications. The results were obtained by averaging over 100 shots.

plasma generation and evolution processes (similar to those described in Ref. [34]). (2) From Fig. 4, a higher incident LE results in a higher NII /Hα ratio; moreover, a higher pressure (density) results in a higher NII /Hα ratio. It is noted that plasma absorption and scattering of the incident laser pulse increase for LEs above the breakdown threshold [40]. The LE dependence of the NII /Hα ratio increases under higher pressure conditions. At elevated pressures, a higher density plasma is generated because of the higher interaction cross section of the laser and the higher concentration of target species near the focal region. Further, the higher loading of the LE at the focus longitudinally expands the plasma well beyond the beam waist. Such a pressure dependency was applied as a new method for gas-density measurement utilizing LIBS in a high-speed environment [7]. (3) In most cases, the line ratios on the left side of the focal zone (− 1.25 to 0 mm) are higher than those on the right side. Since the laser scattering significantly increases beyond the focal point, the energy deposition on the left side is higher than that on the right side, favoring higher NII /Hα ratios at those positions. (4) The variations in the individual emission lines in the Hencken flame are shown to indicate plasma generation and evolution along the laser propagation, which shows that the LIBS plasma is not uniform. However, the plasma has a fixed spatial profile for individual emission lines and NII /Hα ratios at a given pressures. In the next section, we present how the characteristic pressuredependent spatial feature of the LIBS for NII /Hα can be calibrated to perform the 1D-FAR measurement. 3.2. FAR calibration for 1D ns-LIBS line emissions in a Hencken flame at 1–11 bar The goal of FAR calibration is to generate 1D LIBS calibration curves at different pressures. In various spatially uniform Hencken flames at different FARs, spatially resolved LIBS emissions are obtained. The variations in individual emission lines in the Hencken flames are shown to indicate the significant differences in the

plasma generation and evolution processes. The NII /Hα ratio becomes very unstable for 1D measurements at 11 bar, indicating the limitation of the plasma stability, and thus, the FAR calibration methods. Figures 5–7 show the spatial distributions of the NII (568 nm), Hα (656 nm), and NII /Hα line emissions at various equivalence ratios at pressures of 1, 6, and 11 bar, respectively. The relative standard deviations (RSDs) are shown for each set of conditions. In order to quantify the emission signals, the spectral intensity has been calculated from the top 80 % of the intensity profile. 80% of the peak value was used to calculate the N and H emissions, which was to empirically determined to minimize the offset of the continuum background. Because of the influence of collisional broadening, especially under elevated pressure conditions, the instabilities of plasma emission increase [41,42]. According to a previous analysis of the fluctuation in the peak intensities of Hα (656 nm)/NII (568 nm), the optimal LEs for stable LIBS measurements (with a 50-mm-focal-length lens) at pressures of 1, 6, and 11 bar are 150, 100, and 50 mJ/pulse, respectively [34]. A camera gate delay Tdelay = 80 ns and a gate width of 20 ns were used for all pressure conditions. The spectral resolution is about 0.06 nm, which is determined by the spectrometer configuration. The spatial resolution is about 100 μm. The Hα emission has a bimodal appearance in the spatial domain, as shown in the center plots in Figs. 5–7, most prominently at 1 and 6 bar [38]. This feature is also observed in the NII (568 nm) emission, particularly for 11 bar. The laser spark originates from a central point and then propagates both toward and away from the focal point, resulting in the bimodal structure and confirming the plasma evolution. It is noted that the secondary peak of Hα after the focal point is always located around + 0.5 mm, whereas the first peak of Hα before the focal point is moving in the opposite direction of the incident laser beam from − 0.5 mm to − 1.0 mm as the pressure increases from 1 bar to 11 bar. This is because a higher molecular density absorbs a larger amount of LE, thereby causing electron generation from the cascaded ionization processes in the discreet regions before the focal position as the pressure increases. The spatial distributions of the spectral lines yield unique N/H spatial distributions, as shown in subplot (c) of Figs. 5–7. The spatial distributions of the individual N and H intensities and N/H ra-


125

Fig. 6. (Top) Spatial distributions of the mean spectral emission intensities of Hα , NII , and the NII /Hα ratio in methane–air flames at a pressure of 6 bar. (Bottom) The corresponding relative standard deviations (RSDs) are presented. The results were obtained by averaging over 100 shots.

Fig. 7. (Top) Spatial distributions of the mean spectral emission intensity of Hα , NII , and the NII /Hα ratio in methane–air flames at a pressure of 11 bar. (Bottom) The corresponding relative standard deviations (RSDs) are presented. The results were obtained by averaging over 100 shots.

tios are averaged over 100 laser shots with the corresponding RSD. In general, the RSD of the line intensity is inversely proportional to the averaged emission intensity, as illustrated in Fig. 6(d) and (e). However, the RSD of the line-to-line ratio has a much lower correlation with the intensity ratio [i.e., Hα (656 nm)/NII (568 nm) in subplot (f) of Figs. 6 and 7]. As the pressure increases in the elevated combustion environment, the N/H ratio greatly weakens with the loss of a good correlation of line emissions. As shown in Figs. 6 and 7, the RSD of the line ratios increases with the pressure. At 6 bar, the RSDs of both the line ratio and line intensity are approximately the same (∼10 %). The RSD of the line ratio is much worse than the RSD of the line intensities at 11 bar. Clearly, the pressure plays a key role that influences the fluctuations in the line intensity and line-to-line ratios. Under elevated pressure conditions, more absorbers near the focal point give rise to a higher chance for increased seed-electron formation at multiple positions.

The following discrete local energy deposition and absorption in the overall large region significantly impairs further plasma evolution with highly unsteady characteristics (e.g., position, shape, temperature, etc.), leading to large fluctuations in the N/H ratio at 11 bar. As the intensity of laser beam along the 1D line becomes non-uniform, the threshold effect of breakdown becomes very unstable and thus fluctuations in the breakdown lead to large uncertainties in the fuel/air ratio measurements at 11 bar. It is estimated that the uncertainties of the FAR measurements is about 10% for the pressures at 1 and 6 bar, based on the standard variations of the LIBS signals. In addition to the spatial distributions of the line intensities and line ratios (Figs. 5–7), the correlation between the FAR and the line ratio [NII (568 nm)/Hα (656 nm)] has been studied. It is observed that the spatial distributions of the normalized line ratio are identical to an equivalence ratio φ ranging from 0.8 to

126


Fig. 8. Spatial distributions of NII (568 nm) and Hα (656 nm) in premixed laminar methane–air flames at different equivalence ratios at atmospheric pressure. The spatially integrated signals are normalized by the maximum of each curve in (a) and (b). The dashed lines represent the curve fittings.

Fig. 9. (a) Spatially normalized Hα (656 nm)/NII (568 nm) distribution for φ = 1. (b) Correlation between Hα (656 nm)/NII (568 nm) and φ at the focal point (P0 = 0 mm).

1.4. Three representative equivalence ratios, φ = 0.8, 1.0, and 1.2, at pressures of 1, 6, and 11 bar are shown in Figs. 5–7. The line ratio (NII (568 nm)/Hα (656 nm)) decreases as the equivalence ratio increases, as expected. However, the measurement fluctuation (i.e., the RSD) increases with the pressure. A large fluctuation was observed at 11 bar; such subtle discrimination of the line-to-line ratio limits the sensitivity/accuracy of 1D FAR measurement. The RSD in the spatial window of − 0.75 to 0.75 mm is significantly better at 6 bar compared to that at 11 bar. The shapes of the NII /Hα curves for each equivalence ratio are qualitatively identical. For each spatial point, the NII /Hα ratio decreases as the equivalence ratio increases. Therefore, this result is promising for the measurement of the 1D FAR using line ns-LIBS up to 6 bars. The details for the calibrated 1D FAR in premixed, laminar, atmospheric-pressure flames are discussed in Section 4. We also found that the stability of the line ns-LIBS measurement is higher at a lower pressure. For pressures up to 6 bar, the preferable spatial window for the 1D FAR measurement is ± 1.0 mm near the laser focal position with a lens with F = + 50 mm.

tions of the NII (568 nm) and Hα (656 nm) emissions are normalized by the maximum of each curve. Clearly, the identical spatial distributions of NII (568 nm) and Hα (656 nm) confirm that the spatial evolutions of the atomic emissions along the longitudinal direction are independent of φ . However, the amplitude of the N/H ratio decreases with the equivalence ratio, as shown earlier in Figs. 5–7. Furthermore, the bimodal distribution of Hα (656 nm) and unimodal distribution of NII (568 nm) reveal that the difference in the ionization energies of these two elements plays a very important role in determining the spatial distribution of the signal. Figure 9(a) shows the calibration curve for the normalized Hα (656 nm)/NII (568 nm) ratio (HNR) in the spatial domain obtained from the LIBS signal of the flat Hencken flame at φ = 1. As discussed earlier, at each of the spatial points of the flat flame, the amplitude of the N/H ratio has an inverse dependence on the equivalence ratio (although the spatial distribution remains same for all φ ), and the calibration curve in Fig. 9(b) can be used to extract the quantitative 1D FAR from line LIBS for more complex flames. Furthermore, it is found that the linear correlation between the HNR and φ is universally applicable at spatial positions from − 1.0 mm to + 1.0 mm. Note that the linearity of the pointwise HNR with φ is already well-known [34].

4. One-dimensional ns-LIBS application in a turbulent flame In the previous section, the capability of line LIBS for the 1D FAR measurement was demonstrated in a premixed flame under various equivalence-ratio and pressure conditions. Since a singleshot LIBS signal could be obtained under various pressure conditions with a relatively low RSD, we explore 1D ns-LIBS FAR measurement in a turbulent flame at atmospheric pressure. The 1D FAR measurements in turbulent methane–air flames were acquired using a lifted jet flame. For this experiment, a lens with F = + 100 mm was used. With a lens with a longer focal length, the spatial dimension for measurement is expanded. A new set of spatial and spectral calibration curves was obtained for the lens with F = + 100 mm with an input LE of 150 mJ/pulse. The gate delay was set at 80 ns with a camera gate width of 20 ns. 4.1. Calibration for ns-LIBS-based 1D FAR measurements Figure 8 shows the 1D spatial distributions of the NII (568 nm) and Hα (656 nm) emissions for φ = 0.7–1.3. The spatial distribu-

4.2. FAR measurement in a turbulent flame Quantitative 1D-FAR measurements were performed in both laminar and turbulent flames. The spectra of consecutive 100-shot FAR measurements using a 10-Hz laser in laminar and turbulent flames are presented in Figs. 10(a) and 10(b), respectively. The laminar methane–air flame (φ = 1.0) was generated by a Hencken burner with a flow velocity and an outlet Reynolds number of ∼25 cm/s and 400, respectively. The focal position was located at 20 mm above the burner surface. A straight tube with a diameter of 4 mm and length of 20 cm was used to generate a jet turbulent flame. The tube is long enough to be a fully-mixed flow, avoiding nonuniformity of the flow. Since there are no flame holders in the tube flame, the jet turbulent flame (φ ∼ 8.0) was produced with a premixed methane–air flow; the flow velocity was 6 m/s, and the exit Reynolds number was 1600. Note that the equivalence ratio was set high to avoid blowoff. The focal position was set 80 mm


Fig. 10. Single-shot equivalence-ratio measurement results in (a) a laminar flame with φ = 1.0 and (b) a turbulent flame with φ = 8.0 for the initial mixture. The position 0 represents the center of the laser focal point.

above the jet exit. It should be noted that the calibration on the Hencken burner is for equivalence ratios of 0.6–1.6. Beyond these equivalence ratios, the Hencken flame becomes unstable and cannot be used as calibration source. By assuming that the H/N ratios are linear for the equivalence ratios from 0 to 4, an extrapolation of the calibration curve was conducted. Since further fuel–air mixing occurs along the flame, a lower equivalence ratio is observed at this position of Y/D = 12. As the experimental results in Fig. 10(a) show, the overall equivalence ratio in the laminar flame is around 1.0 with a slight fluctuation before the laser focal point (− 1.0 to − 0.3 mm). The fluctuation reaches the minimum value at the focal point. The fluctuation before the focal point is higher than that at the later part because of seedelectron generation; hence, the plasma expansion is higher before the laser beam reaches the focal point. The 1D LIBS profiles in the turbulent flame are quite different from those in the laminar flame and vary significantly from shot to shot. A few selected single-shot line-LIBS results from laminar and turbulent flames are shown in Fig. 11. In contrast to the laminar flame measurements, the local FAR for the turbulent flame significantly varies at different spatial locations. Although the initial equivalence ratio was set at 8.0 at the outlet of the tube burner, the surrounding entrained air was continuously transported and dispersed into the original fuel–air mixture by turbulent motion. At a height of 80 mm (Y/D = 20), the acquired equivalence ratio was ∼2.5 with the spatial region of − 1.0 to + 1.0 mm.

127

Fig. 11. Single-shot equivalence ratio measurements in both laminar- and turbulent-flame environments extracted from a few individual laser shots selected from Figs. 10(a) and 10(b). The fluctuation reaches the minimum value at the focal point. Moreover, the fluctuation before the focal point is higher than that after the focal point.

Fig. 12. Experimental setup for proof-of-principle multiline LIBS.

5. Potential of multiline LIBS for Quasi-2D measurement In this section, we study the feasibility of using multiline LIBS for larger-area FAR measurements. As shown in Fig. 12, proof-ofprinciple multiline LIBS is demonstrated with a homemade linear lens array. This linear lens array was composed of five 6mm-diameter spherical lenses with F = + 100-mm placed in parallel. The 8-mm-diameter laser beam from a Pro350 Nd:YAG laser with an energy of ∼1 J/pulse was conditioned to form a 20-mmtall and 6-mm-thick collimated beam using four cylindrical lenses. The beam was then focused using a linear lens array to generate five plasma lines. An ICCD camera with a stereoscope was used to acquire the Hα (656 nm) and NII (568 nm) emissions simultaneously. Two single-band bandpass filters with center wavelengths of 568 and 656 nm were employed in front of the imaging system to

Fig. 13. Line images of H and N emission from multiline LIBS at various FARs in atmospheric flames.

collect the NII and Hα line emissions, respectively. It was observed that the emission intensity of the middle line is stronger than that of the side line because the input laser beam has a Gaussian distribution, where the center part of the beam has a higher energy. The LEs for the five lines (from top to bottom) are 53, 85, 95, 70, and 58 mJ/pulse. FAR calibration for multiline LIBS was conducted in premixed methane–air flames from a Hencken burner. Figure 13 shows the image profiles of H and N emission lines at equivalence ratios of 0.7–1.3 in the Hencken flames. The laser

128


Fig. 14. Normalized 1D spatial profiles of H (left top) and N (left bottom) emissions at various equivalence ratios of the centerline (Pos_3). (Right) Linear correlation between the H/N ratio and the equivalence ratio for the center of the profile (circled points).

beam was split into five lines. Figure 13 shows the possibilities of extending to multiline measurements of the FAR in the flame. Owing to the variations in the energies of different laser beams, the absolute intensities of individual lines are very different. The individual line intensities are not correlated with the FAR. The image profile of the center line (Pos_3) is shown in Fig. 14. Similar to those presented in Figs. 5 and 8, the normalized emissions of N and H lines remain the same for the entire range of equivalence ratios, indicating that the spatial distribution of the entire line remains unchanged from fuel-lean to fuel-rich conditions. This is consistent with the single-line measurement. A good linear correlation between the H/N ratio and the equivalence ratio is achieved at the test position. 6. Summary and discussion Line LIBS for spatially resolved 1D FAR measurements in laminar and turbulent flames is demonstrated. There are three major conclusions of the results presented here. (1) ns-LIBS 1D FAR measurements in laminar methane–air flames were studied at pressures varying from 1 bar to 11 bar. (2) With careful calibration of the spatial distribution of the measured H/N ratio, line LIBS was successfully employed for 1D FAR measurements in both laminar and turbulent flames at atmospheric pressure. (3) A proof-of-principle extension to multiline FAR measurement was demonstrated. In all of the above cases, it was observed that the spatial distributions of the LIBS signals from NII and Hα do not change with the initial equivalence ratios. Such an invariance in the spatial distribu-

tion enables the spatial calibration of the LIBS signal for FAR measurements, which can be applied to laminar and turbulent flames to determine the spatially resolved FAR. A local equivalence-ratio distribution with a spatial resolution of 0.1 mm was achieved. The uncertainty in the measurement of the 1D FAR was less than 10 % (within the 2-mm probe volume) at 1 bar. The uncertainty increases with pressure. Proof-of-principle multiline LIBS measurements of the local FAR show promise for extending FAR measurements into a large region of interest. The drawbacks of 1D ns-LIBS for fuel/air ratio measurements include limited spatial domain up to a few millimeters and a more complex calibration procedure than traditional LIBS. A high-power femtosecond laser might be able to extend the spatial domains up to a few centimeters, although ratios of emission lines of H and O will have to be used. [43] Funding Air Force SBIR: FA8650-15-M-2593; Air Force Research Laboratory (AFRL): FA8650-16-C-2725; FA8650-15-D-2518; Air Force Office of Scientific Research (AFOSR) LRIR: 14RQ06COR; NSF 1,418,848. Acknowledgment This document has been approved for public release by AFRL (Distribution A, No. 88ABW-2018-2312). References [1] Y.P. Raizer, Gas discharge physics, Springer, New York, 1997. [2] D.C. Smith, R.T. Brown, Aerosol-Induced air breakdown with Co2 -laser radiation, J. Appl. Phys. 46 (3) (1975) 1146–1154.

Y. Wu et al. / Combustion and Flame 198 (2018) 120–129 [3] A.A. Boni, D.A. Meskan, Effect of impurities on laser-induced air breakdown at 1.06mu, Opt. Commun. 14 (1) (1975) 115–118. [4] J.L. Beduneau, Y. Ikeda, Spatial characterization of laser-induced sparks in air, J. Q. Spectr. Radiat. Transfer 84 (2) (2004) 123–139. [5] Y. Wu, J.C. Sawyer, L. Su, Z. Zhang, Quantitative measurement of electron number in nanosecond and picosecond laser-induced air breakdown, J. Appl. Phys. 119 (17) (2016) 173303. [6] H. Do, C. Carter, Hydrocarbon fuel concentration measurement in reacting flows using short-gated emission spectra of laser induced plasma, Combust. Flame 160 (3) (2013) 601–609. [7] H. Do, C.D. Carter, Q.L. Liu, T.M. Ombrello, S. Hammack, T. Lee, K.Y. Hsu, Simultaneous gas density and fuel concentration measurements in a supersonic combustor using laser induced breakdown, Proc. Combust. Inst. 35 (2015) 2155–2162. [8] N. Kawahara, E. Tomita, S. Takemoto, Y. Ikeda, Fuel concentration measurement of premixed mixture using spark-induced breakdown spectroscopy, Spectrochim. Acta Part B-Atom. Spectr. 64 (10) (2009) 1085–1092. [9] C. Letty, A. Pastore, E. Mastorakos, R. Balachandran, S. Couris, Comparison of electrical and laser spark emission spectroscopy for fuel concentration measurements, Exp. Thermal Fluid Sci. 34 (3) (2010) 338–345. [10] B. McGann, C.D. Carter, T. Ombrello, H. Do, Direct spectrum matching of laser-induced breakdown for concentration and gas density measurements in turbulent reacting flows, Combust. Flame 162 (12) (2015) 4479–4485. [11] M.T.A. Rahman, N. Kawahara, K. Tsuboi, E. Tomita, Effect of ambient pressure on local concentration measurement of transient hydrogen jet in a constant-volume vessel using spark-induced breakdown spectroscopy, Int. J. Hydrogen Energy 40 (13) (2015) 4717–4725. [12] T.W. Lee, N. Hegde, Laser-induced breakdown spectroscopy for in situ diagnostics of combustion parameters including temperature, Combust. Flame 142 (3) (2005) 314–316. [13] J. Kiefer, J.W. Troger, Z.S. Li, T. Seeger, M. Alden, A. Leipertz, Laser-induced breakdown flame thermometry, Combust. Flame 159 (12) (2012) 3576–3582. [14] T.X. Phuoc, F.P. White, Laser-induced spark for measurements of the fuel-to-air ratio of a combustible mixture, Fuel 81 (13) (2002) 1761–1765. [15] F. Ferioli, P.V. Puzinauskas, S.G. Buckley, Laser-induced breakdown spectroscopy for on-line engine equivalence ratio measurements, Appl. Spectr. 57 (9) (2003) 1183–1189. [16] P. Stavropoulos, A. Michalakou, G. Skevis, S. Couris, Quantitative local equivalence ratio determination in laminar premixed methane-air flames by laser induced breakdown spectroscopy (LIBS), Chem. Phys. Lett. 404 (4-6) (2005) 309–314. [17] P. Stavropoulos, A. Michalakou, G. Skevis, S. Couris, Laser-induced breakdown spectroscopy as an analytical tool for equivalence ratio measurement in methane-air premixed flames, Spectrochim. Acta Part B-Atom. Spectr. 60 (7-8) (2005) 1092–1097. [18] F. Ferioli, S.G. Buckley, Measurements of hydrocarbons using laser-induced breakdown spectroscopy, Combust. Flame 144 (3) (2006) 435–447. [19] L. Zimmer, S. Tachibana, Laser induced plasma spectroscopy for local equivalence ratio measurements in an oscillating combustion environment, Proc. Combust. Inst. 31 (2007) 737–745. [20] S. Joshi, D.B. Olsen, C. Dumitrescu, P.V. Puzinauskas, A.P. Yalin, Laser-Induced Breakdown Spectroscopy for In-Cylinder Equivalence Ratio Measurements in Laser-Ignited Natural Gas Engines, Appl. Spectr. 63 (5) (2009) 549–554. [21] M.S. Mansour, H. Imam, K.A. Elsayed, W. Abbass, Local equivalence ratio measurements in turbulent partially premixed flames using laser-induced breakdown spectroscopy, Spectrochim. Acta Part B-Atom. Spectr. 64 (10) (2009) 1079–1084. [22] S.B. Gupta, B. Bihari, M. Biruduganti, R. Sekar, In-cylinder equivalence ratio measurements in an EGR equipped engine, J. Eng. Gas Turb. Power-Trans. ASME 133 (11) (2011) 114504-1–114504-5.

129

[23] S.H. Zhang, X.L. Yu, F. Li, G.J. Kang, L.H. Chen, X.Y. Zhang, Laser induced breakdown spectroscopy for local equivalence ratio measurement of kerosene/air mixture at elevated pressure, Opt. Lasers Eng. 50 (6) (2012) 877–882. [24] M.M. Tripathi, K.K. Srinivasan, S.R. Krishnan, F.Y. Yueh, J.P. Singh, A comparison of multivariate LIBS and chemiluminescence-based local equivalence ratio measurements in premixed atmospheric methane-air flames, Fuel 106 (2013) 318–326. [25] J. Kiefer, B. Zhou, Z.S. Li, M. Alden, Impact of plasma dynamics on equivalence ratio measurements by laser-induced breakdown spectroscopy, Appl. Opt. 54 (13) (2015) 4221–4226. [26] T.X. Phuoc, Laser-induced spark ignition fundamental and applications, Opt. Lasers Eng. 44 (5) (2006) 351–397. [27] M. Kotzagianni, E. Kakkava, S. Couris, Laser-Induced Breakdown Spectroscopy (LIBS) for the Measurement of Spatial Structures and Fuel Distribution in Flames, Appl. Spectr. 70 (4) (2016) 627–634. [28] T.X. Phuoc, Laser-induced spark for simultaneous ignition and fuel-to-air ratio measurements, Opt. Lasers Eng. 44 (6) (2006) 520–534. [29] A. Michalakou, P. Stavropoulos, S. Couris, Laser-induced breakdown spectroscopy in reactive flows of hydrocarbon-air mixtures, Appl. Phys. Lett. 92 (8) (2008) 081501. [30] S.B. Han, Y.J. Chung, The influence of air-fuel ratio on combustion stability of a gasoline engine at idle, SAE International, 1999. [31] L. Zimmer, K. Okai, Y. Kurosawa, Combined laser induced ignition and plasma spectroscopy: Fundamentals and application to a hydrogen-air combustor, Spectrochim. Acta Part B-Atom. Spectr. 62 (12) (2007) 1484–1495. [32] M. Buschbeck, F. Buchler, T. Halfmann, S. Arndt, Laser-induced breakdown spectroscopy for lambda quantification in a direct-injection engine, Spectrochim. Acta Part B-Atom. Spectr. 74-75 (2012) 103–108. [33] J. Kiefer, J.W. Troger, Z.S. Li, M. Alden, Laser-induced plasma in methane and dimethyl ether for flame ignition and combustion diagnostics, Appl. Phys. B-Lasers Opt. 103 (1) (2011) 229–236. [34] P.S. Hsu, M. Gragston, Y. Wu, Z.L. Zhang, A.K. Patnaik, J. Kiefer, S. Roy, J.R. Gord, Sensitivity, stability, and precision of quantitative Ns-LIBS-based fuel-air-ratio measurements for methane-air flames at 1-11bar, Appl. Opt. 55 (28) (2016) 8042–8048. [35] M. Kotzagianni, R. Yuan, E. Mastorakos, S. Couris, Laser-induced breakdown spectroscopy measurements of mean mixture fraction in turbulent methane flames with a novel calibration scheme, Combust. Flame 167 (2016) 72–85. [36] S. Couris, M. Kotzagianni, A. Baskevicius, T. Bartulevicius, V. Sirutkaitis, Combustion diagnostics with femtosecond laser radiation, 22nd International Conference on Spectral Line Shapes (ICSLS 2014) (2014), p. 548. [37] Y.L. Chen, J.W.L. Lewis, C. Parigger, Spatial and temporal profiles of pulsed laser-induced air plasma emissions, J. Q. Spectr. Radiat. Transfer 67 (2) (20 0 0) 91–103. [38] N. Glumac, G. Elliott, The effect of ambient pressure on laser-induced plasmas in air, Opt. Lasers Eng. 45 (1) (2007) 27–35. [39] A. Kramida, Yu. Ralchenko, J. Reader, NIST ASD Team, NIST atomic spectra database (version 5.4), [Online], in Mon Dec 19 2016. 2016, NIST. [40] Z. Zhang, M.N. Shneider, R.B. Miles, Coherent microwave Rayleigh scattering from resonance-enhanced multiphoton ionization in argon, Phys. Rev. Lett. 98 (26) (2007). [41] D.A. Cremers, A.K. Knight, Laser-induced breakdown spectroscopy, Wiley Online Library, 2006. [42] P.S. Hsu, M. Gragston, Y. Wu, Z. Zhang, A.K. Patnaik, J. Kiefer, S. Roy, J.R. Gord, Sensitivity, stability, and precision of quantitative Ns-LIBS-based fuel-air-ratio measurements for methane-air flames at 1–11bar, Appl. Opt. 55 (28) (2016) 8042–8048. [43] A.K. Patnaik, P.S. Hsu, Y. Wu, M. Gragston, Z. Zhang, J.R. Gord, S. Roy, Ultrafast Laser-Induced-Breakdown Spectroscopy (LIBS) for F/A-ratio measurement of hydrocarbon flames, AIAA Aerospace Sciences Meeting, American Institute of Aeronautics and Astronautics (2018).