2019ApJ...887...46Z

The PAH emission in Spitzer-IRS spectral maps of the reflection nebula NGC 2023 have been previously studied using a Gaussian decomposition method for the 7-9 µm region and a database-fitting approach. Both studies provided insight into the spatial-spectral evolution of the PAH population and related them to changing local physical conditions. This study investigates whether the database-fitting technique provides insight into the PAH populations at the origin of the four Gaussian components. To this end, clustered PAH species maps and spectra are generated from the database-fitting results using spectral clustering utilizing the Structural Similarity Index as an affinity measure. The application of spectral clustering solely based on spatial structure is strongly dependent on the anatomy of the considered regions and is unable to align specific morphological features with a PAH population characterized by a single distinct property. However, in the south FOV the projected distance from the star of the peak emission in a cluster map correlates with the PAH cation fraction and the cluster dominated by small PAHs is confined to the S and SSE ridges, consistent with results from Knight et al. Furthermore, the cluster and Gaussian maps exhibit limited morphological similarity and the 7-9 µm cluster spectra do not show consistent overlap with any of the Gaussian components. However, the relative strengths of the Gaussian components strongly correlate with the PAH ionization parameter as determined from the database-fitting approach. This lends further support to the existence of at least two sub-populations contributing to the 7-9 µm PAH emission.