📝 SummarXiv

Abstract

A galaxy's mid-IR spectrum encodes key information on its radiation field, star formation, and dust properties. Characterizing this spectrum therefore offers strong constraints on a galaxy's activity. This project describes a diagnostic tool for identifying main-sequence (MS) star-forming galaxies (SFGs) in the local Universe using IR dust emission features that are characteristic of galaxy activity. A physically-motivated sample of mock galaxy spectra has been generated to simulate the IR emission of SFGs. Using this sample, we developed a diagnostic tool for identifying MS SFGs based on machine learning methods. Custom photometric bands were defined to target dust emission features, including polycyclic aromatic hydrocarbons (PAHs) and the dust continuum. Three bands were chosen to trace PAH features at 6.2 {\mu}m, 7.7 {\mu}m, 8.6 {\mu}m, and 11.3 {\mu}m, along with an additional band to probe the radiation field strength responsible for heating the dust. This diagnostic was subsequently applied to observed galaxies to evaluate its effectiveness in real-world applications. Our diagnostic achieves high performance, with an accuracy of 90.9% on MS SFGs (observed sample of SFGs). Additionally, it shows low contamination, with only 16.2% of AGN galaxies being misidentified as SF. Combining observational data with stellar population synthesis models enables the creation of physically-motivated samples of SFGs that match the spectral properties of real galaxies. By positioning custom photometric bands targeting key dust features, our diagnostic can extract valuable information without the need to measure emission lines. Although PAHs are sensitive indicators of star formation and interstellar medium radiation hardness, PAH emission alone is insufficient for identifying MS SFGs. Finally, we developed a physically-motivated spectral library of MS SFGs spanning from UV to FIR wavelengths.