Baseline-free wavelength modulation spectroscopy based on cepstral analysis

This paper demonstrates what we believe to be a new WMS m-FID technique based on cepstral analysis, incorporating wavelength modulation spectroscopy (WMS) with a modified form of the time-domain molecular free-induction decay (m-FID) signal. Detailed theoretical framework of the WMS m-FID technique, as well as the fitting routine, has been investigated. The proposed WMS m-FID technique is first validated through a static CO gas cell experiment using the CO <i>R</i>(9) absorption line near 2179.7719 cm^-1. The fitting error between the measured and best-fit m-FID signals is less than 1.0 <i>%</i>, with integrated absorbance area of relative uncertainty 0.17 <i>%</i>. Time-resolved measurements of CO over a 6-second purge period demonstrate the reliability and robustness of the proposed WMS m-FID approach in comparison with direct absorption spectroscopy (DAS) and WMS-2<i>f</i>/1<i>f</i> techniques. The demonstration of the proposed WMS m-FID approach is further carried out in a premixed CH₄/Air laminar flame of a flat-flame burner. The relative fitting residuals between the measured and best-fit m-FID signals remain below 4.0 <i>%</i>, yielding <i>T</i> = 1763 K and <i>X</i>_H₂O=16.58 <i>%</i> at HAB = 7 mm, with a temperature uncertainty of 30 K and absolute H₂O concentration uncertainty of 0.65 <i>%</i>. The average temperature difference between the WMS m-FID and WMS-2<i>f</i>/1<i>f</i> techniques is within 18.0 K (<1.0 <i>%</i> relative), and the H₂O concentration difference is within 0.4 <i>%</i> (<2.6 <i>%</i> relative). Additionally, the proposed WMS m-FID technique significantly improves computational efficiency, enabling 22-fold acceleration compared to the time-consuming spectral-fitting routine in standard WMS-2<i>f</i>/1<i>f</i> techniques. Finally, a strictly baseline-free CO measurement is achieved by directly fitting the m-FID signal of the measured transmitted intensity beyond the 3^rd-order peak (<i>τ</i> ≥ ~5.7 ns) and the 4^th-order peak (<i>τ</i> ≥ ~7.7 ns), with deviations of 2.7 <i>%</i> and 4.4 <i>%</i> from the entire m-FID signal fitting, respectively. Both the static gas-cell and high-temperature flat-flame demonstrations indicate that the proposed WMS m-FID technique enables quantitative, accurate, and baseline-free measurements. It is envisioned that the WMS m-FID technique can be an effective solution for achieving baseline-free, quantitative, <i>in situ</i> monitoring in practical combustion environments.