In the mid-infrared (3–20 μm), our universe is dominated by broad spectral features. These features are though to be due to large carbonaceous molecules called Polycyclic Aromatic Hydrocarbons (PAHs). In the 10 to 15 μm range, several of these bands show variations in shape and intensity depending on the astronomical environments where they are seen. The exact origin of these variations is not explained. In this paper, we apply two different methods to hyper-spectral data to help analyze these variations: decomposition into end-members using non-negative matrix factorization and linear fitting using theoretical database spectra. Using both methods, we find that three chemical populations of PAHs are needed. The first two populations: PAHs in the neutral and ionized states, are identified by the two methods. The variation in relative abundance between these two chemical populations is shown to be responsible for most of the observed spectral variations. However, the third component found by NMF is inconsistent with the third component found from the theoretical database fitting. We discuss the limitations and benefits of each method for the analysis of astronomical spectra.
展开▼
