Extracting ion emission lines from femtosecond-laser plasma x-ray spectra heavily contaminated by spikes

Nowadays charged-coupled device (CCD) detectors are widely used for the registration of multicharged ions x-ray spectra. These spectra are generated in a plasma during interaction of ultrashort, ultraintense laser pulses with solid targets. Strong parasitic radiation from the plasma affects CCD detectors and contaminates resulting spectra, so that spectral features can be completely covered by noise even during measurements with a very short accumulation time. In this work we propose a "mean to median" (M2M) algorithm for noise suppression in femtosecond laser plasma x-ray spectra. Series of spectra is necessary for the identification of corrupted data points by the developed method. The algorithm was tested with model spectra which reflect main features of experimental data. In practice we used it for extracting information about spectral lines of Ne-like Fe ions and He-like Al ions which allowed us to calculate plasma parameters. It is demonstrated that M2M method is able to clean spectra with more than 10% of corrupted pixels. Fluctuations in intensity of spectral lines induced by laser instability do not affect validity of the proposed method.