Presentation #130.03 in the session Star Associations.
Recent studies of the Milky Way suggest that extended, coherent filaments of dense gas form the backbone of the cold interstellar medium in galaxies. Extinction mapping with Hubble Space Telescope (HST) offers almost the only way to achieve the < 10 pc resolution needed to identify dense filamentary structures in galaxies beyond the Local Group. In a new archival HST program, we have combined Hubble’s outstanding resolution and color information with machine learning (convolutional neural network) segmentation methods and image inpainting (masked pixel restoration) techniques to systematically identify/characterize coherent dust features and quantify the distribution of extinction across the PHANGS-HST sample of 38 nearby galaxies (each with NUV-U-B-V-I imaging). All our targets also have ALMA CO imaging, though at notably lower resolution than achieved by HST. We present the initial results of this archival study. Our analysis will allow us to explore using the detailed structure of extinction as a high-resolution proxy of the less-resolved molecular gas distribution and compare the overall dust and molecular morphology on much larger galaxy-wide scales. We also will directly compare our observed extinction maps to synthetic observations from leading theoretical simulations. Our work will provide a strong test of whether this new view of the Milky Way’s ISM is a general feature of galaxies or restricted to specific environments like spiral arms or stellar bars.