Investigating supernova-remnant X-ray maps with dendrogram-based techniques

X-ray images of supernova remnants provide important information regarding the spatial distribution of the physical and chemical parameters of the shocked plasma. Our objective is to provide a comprehensive and quantitative assessment of the similarities observed across narrow-band images and equivalent-width maps of a given supernova remnant. We also aim to perform comparative analyses between distinct remnants. We adopted two dendrogram-based techniques, namely a multivariate analysis and a deep-learning, graph-based approach (by exploring different metrics) for the study of Chandra X-ray maps of Cassiopeia A, G292.0+1.8, Tycho's SNR, Kepler's SNR, and W49B. We also tested the reliability of our results on synthetic X-ray maps produced from a 3D hydrodynamic simulation of Cassiopeia A. We identify clusters of narrow-band images (and equivalent-width maps) corresponding to emission lines of chemical elements that are expected to be synthesized in the same layers of the exploding star and/or that originate from plasma at similar temperatures. By comparing different objects, we clearly observe strong similarities among Type Ia remnants (namely Tycho's SNR and Kepler's SNR) and core-collapse remnants (Cassiopeia A and G292.0+1.8), with W49B being more similar to the latter objects. The methodology developed in this work provides a robust tool to measure the analogies and differences between narrow-band images and equivalent-width maps of a supernova remnant, offering clues on the distribution of the different chemical species and on the plasma temperature. Although tested on a limited sample of objects, the adoption of this tool for the classification of supernova remnants (Type Ia vs. core-collapse) yields promising results.