Direct spectrum matching of laser-induced breakdown for concentration and gas density measurements in turbulent reacting flows

Abstract

A direct spectrum matching method for laser-induced breakdown spectroscopy is proposed to simultaneously measure gas density and concentration in turbulent reacting environments with improved measurement accuracy. The breakdown spectrum recorded in the target flow is directly matched with a spectrum out of a database consisting of various emission spectra recorded under well-defined conditions in a range of gas density and composition. It is shown that the wavelength, intensity and line width of the atom/ion emission lines in the spectrum indicate atom composition and gas density that are independent of parent molecular species in the target flow. Once a matching spectrum (within 550–830 nm containing O, H, N, and C lines) in the database of a known gas condition is found, the concentration and gas density at the location of the breakdown can be accurately derived. A 532-nm Nd:YAG laser with 10-Hz pulse repetition rate is used to induce breakdown in fuel/air mixtures in a variable pressure combustion chamber to build the spectrum database.In addition, it is used in a cavity flameholder of a model supersonic combustor to measure the gas density and concentration fields in a turbulent reacting environment. All the measurements are completed within 100 ns after laser firing, before breakdown affects the flow and the fast evolving environment alters the breakdown spectrum.

Introduction

Laser-induced breakdown spectroscopy (LIBS) is a technique that allows for quantitative measurement of gas properties in challenging environments [1], [2], [3]. A laser beam focused with sufficient energy to ionize, dissociate, and excite molecules in a small volume about the focal point creates a small volume of plasma, typically less than 1 mm³ [1], [2], [3], [4], [5], [6], [7], [8]. Electrons of the gas constituents near the focus absorb the photon energy through the inverse Bremsstrahlung process to cause excitation, dissociation and ionization [9]. The prominent mechanism of the laser-induced plasma development is assumed to be through cascade ionization as opposed to multiphoton ionization due to the relationship between gas density and laser energy required to induce breakdown, [10]. Then, the excited species in the laser-induced plasma (e.g., atoms and ions) emit photons containing information on atom composition and gas density. Atom composition in the breakdown plasma determines the wavelength and intensity of the photon emission with individual emission line strength being a function of the atom/ion concentration. In addition, the unperturbed temperature and pressure, i.e., the gas density, at the focal point where the plasma is produced before breakdown, has a great effect on the line width and overall emission strength, since the gas density determines the number of molecules absorbing the laser photons and the number of atoms that emit photons. (Note that throughout this manuscript gas density refers to the ratio of pressure to temperature prior to the formation of the plasma.) The quantitative LIBS measurement is based on the monotonic correlations of the quantities that are extracted from the emission spectra (e.g., emission line widths, strengths and their ratios) and the physical gas properties (e.g., gas density and concentration) of target media.

Implementing the measurements in rapidly evolving turbulent reacting environments, however, is challenging because the breakdown plasma, surrounding high-speed turbulent flows, and combustion chemistry interact to increase measurement uncertainty [11], [12], [13], [14], [15]. In addition, large and rapid variations of gas properties such as density, species concentration, pressure, and temperature in the harsh environment make the measurement difficult. Therefore, a method for deriving gas properties within the shortest time is essential such that the measurement is not affected by the interaction of the breakdown with the surroundings. In previous work [2], [3], a nanosecond-gated LIBS (n-LIBS) method was developed to instantaneously measure species concentration in a reacting flow. This newly developed n-LIBS technique minimizes the total measurement time below 100 ns (versus > 1 μs in conventional LIBS [4], [5], [6], [7], [8]) enabling the application of the technique to turbulent supersonic flows [2], [3]. Density dependence of the n-LIBS is also investigated and is critical for measurements in compressible and/or reacting environments. Previously, several atomic emission line intensities and their ratios were used as the species/atom concentration indicators [1], [2], [3], [4], [5], [6], [7], [8]. However, it was found that the ratio of emission line intensities is also a function of gas density, which limits the application of the method. Furthermore, not only do the emission line intensities (and their ratios) change with gas density, but also the line widths and spectrum baseline profiles are sensitive to the density variation. Therefore, simply extracting the line intensity from the spectra is not the best approach vis-a-vis measurement accuracy.

In this study, we are proposing a Direct Spectrum Matching (DSM) method that accounts for all the detailed spectrum features including the multiple emission line intensities/widths (more than 15 atomic/ionic emission lines), overlapping line structures, baseline profiles, etc. With the DSM method, the goal is to find the best match of the spectrum from a database containing spectra collected at well-defined gas conditions. The DSM process is just like the fingerprint matching process, thus requiring an intensive spectral database. Therefore, the contribution from each atom to the spectrum profile (e.g., sensitivity of line strength to atom concentration) is quantitatively evaluated in this study. It is used to explore the possibility of accurate and empirical spectrum profile prediction that is potentially filling the gaps between the data points in the database and extending the property measurement range. In order to provide the proof of the concept, several sample spectra recorded within a cavity flameholder (in a Mach-2 crossflow) are used to demonstrate the feasibility of the technique in practice.