Processing of Thomson scattering spectra for diagnostics of laser-induced plasma

Abstract

Knowledge of electron temperature and electron number density of laser-induced plasma (LIP) is needed for both fundamental and applied researches. Typically, these parameters are derived from plasma emission spectra. It can cause uncertainties due to plasma inhomogeneity and self-absorption of emission lines. Another approach for diagnostics of LIP supposes registration and further analysis of Thomson scattering (TS) spectra. However, it is also a challenging way due to high level of spectral interferences (plasma emission, Rayleigh and Mie scatterings) from such a low-temperature plasma and low signal-to-noise ratio. The present paper evaluates the possibility of preprocessing of Thomson scattering spectra by wavelet denoising and suggests the approach of LIP diagnostics based on TS spectra. We used numerical simulations to select wavelets appropriate for retrieving TS signal from noisy spectra. The results indicate that wavelet denoising outperforms FFT filtration and provides high-quality results down to extremely small values of signal-to-noise ratio. Also, it was shown that wavelets efficiency varies for different values of TS parameter α, but some of them with “Sym 6” as the best representative are more universal that was proved by processing of experimental spectra.We describe two approaches of plasma parameters determination by numerical fitting of electron features of TS with Salpeter function or by fitting them with Lorentzian profile with further use of calculated nomograms. The first approach can successfully work with TS spectra in the absence of intense Rayleigh scattering. Nomograms approach demonstrated reliable results even in the presence of strong spectral interferences. Therefore, we consider wavelet denoising (Sym 6, Rbio1.3 and Bior 2.2) combined with nomograms approach for parameters calculation to be the most reliable and robust way of LIP diagnostics by TS.