Comparative analysis of fluorescence removal approaches for instantaneous Raman/Rayleigh measurements in NH₃/H₂-air, CH₄-air, and NH₃/CH₄-air flames

This work compares two distinct methodologies for the removal of fluorescence contributions from Raman spectra collected in NH₃/H₂, CH₄, and NH₃/CH₄ counterflow premixed and non-premixed flames. The first approach (1-component) combines instantaneous measurements in a Raman-free spectral region and temperature-dependent correction factors calibrated in a single laminar flame, to estimate fluorescence contribution on the Raman signal. The second approach (2-component) requires the simultaneous acquisition of the vertically and horizontally polarized signal to separate the strongly polarized Raman signal from the polarization independent fluorescence. The accuracy, precision, and limitations of each method are thoroughly assessed under various flame conditions. The 1-component approach provides the same accuracy and higher precision with the 2-component approach in NH₃/H₂ flames where the majority of the fluorescence signal is attributed to a single radical, NH₂. The accuracy of measurements based on the 1-component approach drops for CH₄ and NH₃/CH₄ flames as the composition of the pool of radicals and soot precursors responsible for the fluorescence signal is flame-dependent and a single flame calibration cannot provide general results. In contrast, the polarization-separation approach exhibits inherent robustness by not requiring prior knowledge of fluorescence signal sources. This approach takes into account the depolarization ratio of each molecule in the Raman response curves. For N₂, O₂, H₂, and H₂O, the library is constructed based on the difference between polarized and depolarized signals. For NH₃, the Raman response curve is experimentally calibrated using the difference between these two signals. Building upon previous work, this approach benefits from an improved experimental setup and incorporates the wavelet adaptive thresholding and reconstruction (WATR) denoising algorithm to enhance measurement precision. Successfully applied to NH₃/H₂-air and NH₃/CH₄ flames, the polarization-separation method demonstrates satisfactory accuracy (∼ 0.01 mole fractions, 40 K temperatures) and precision (∼ 0.02 mole fractions, 60 K temperatures) across diverse conditions, regardless of the source of laser-induced fluorescence signal (NH₂, NO₂, C₂, PAH, and other soot precursors).