Separating spectral components for high energy astronomy.
The accepted version of the methodology paper can be found on arxiv here.
The animation belows shows energy slices of a data cube (RA,DEC,E) generated from several deep observations of the CasA SNR as seen by the Chandra X-ray telescope:
The X-ray emission is composed of several components like bremsstrahlung (from the hot gas), synchrotron emission (from particles accelerated at the shock) and a population of emission lines from Si, Mg, Ar, Fe, etc.
The plot below shows a spectrum from the entire SNR and a simple best-fit model.
Producing a map of the emission from individuals lines can be difficult because of components mixing:
- several lines are very nearby in energy
- most lines are dominated by continuum emission and its base level cannot be simply derived from the image
We here propose to use the method of blind (or semi-blind) source separation to separate the spectral components in X-rays. In particular we want to apply the recent methods (LGMCA ; Local-Generalized Morphological Component Analysis) developped by the CosmoStat group at CEA that have been succesfuly applied to Planck to separated the CMB map from the foregrounds emissions (see here for more details).