A machine-learning photometric classifier for massive stars in nearby galaxies II. The catalog

TL;DR

This work combines Spitzer photometry and Pan-STARRS1 optical data to classify point sources in 26 galaxies within 5 Mpc, and provides the largest spectroscopically confirmed catalog of extragalactic massive stars and candidates to date, beyond the Clouds, comprising 5,273 sources.

Abstract

Mass loss is a key aspect of stellar evolution, particularly in evolved massive stars, yet episodic mass loss remains poorly understood. To investigate this, we need evolved massive stellar populations across various galactic environments. However, spectral classifications are challenging to obtain in large numbers, especially for distant galaxies. We addressed this by leveraging machine-learning techniques. We combined Spitzer photometry and Pan-STARRS1 optical data to classify point sources in 26 galaxies within 5 Mpc, and a metallicity range 0.07-1.36 Z$_\odot$. Gaia data release 3 (DR3) astrometry was used to remove foreground sources. Classifications are derived using a machine-learning model developed in our previous work. We report classifications for 1,147,650 sources, with 276,657 sources (~24%) being robust. Among these are 120,479 red supergiants (RSGs; ~11%). The classifier performs well even at low metallicities (~0.1 Z$_\odot$) and distances under 1.5 Mpc, with a slight decrease in accuracy beyond ~3 Mpc due to Spitzer's resolution limits. We also identified 21 luminous RSGs (log($L/L_\odot)\ge5.5$), 159 dusty yellow hypergiants in M31 and M33, as well as 6 extreme RSGs (log($L/L_\odot)\ge6$) in M31, challenging observed luminosity limits. Class trends with metallicity align with expectations, although biases exist. This catalog serves as a valuable resource for individual-object studies and James Webb Space Telescope target selection. It enables the follow-up on luminous RSGs and yellow hypergiants to refine our understanding of their evolutionary pathways. Additionally, we provide the largest spectroscopically confirmed catalog of extragalactic massive stars and candidates to date, beyond the Clouds, comprising 5,273 sources (including ~330 other objects).

Figures (6)

• Figure 1: Example of fitting processing and foreground removal in NGC 2403. Top panel: Gaia sources where green ellipse defines boundary selected for galaxy. Middle panel: Fitting parallax data for all sources within NGC 2403. The foreground contribution (dotted dark gray line) is scaled according to the densities of sources inside and outside the galaxy, while the dashed purple line shows the (Gaussian) contribution of galactic sources, and the blue line indicates the total fit. The vertical dashed black lines show the $3\sigma$ limits defined by the Gaussian properties of the galactic sources. Bottom panel: Similar to parallax, but for proper motion in declination (see Sect. \ref{['s:gaia']} for more details).
• Figure 2: Example of CMDs ($z$ vs. $r-z$ and [4.5] vs. $[3.6]-[4.5]$) for M31 (top) and NGC 2403 (bottom) with predicted classifications. We plot all sources that satisfy the quality cuts imposed in Sect. \ref{['s:quality_cuts']}, and their photometric values come from the original data (i.e., we do not plot sources whose values have been imputed during the application of the classifier). We use cyan circles for blue supergiants (BSG), blue diamonds for B[e] supergiants (BeBR), gray triangles for galaxies (GAL), green pentagons for luminous blue variables (LBV), red crosses for red supergiants (RSG), purple empty triangles for Wolf-Rayet stars (WR), and yellow filled bottom-sided for yellow supergiants (YSG). The total number of sources per class is provided next to the class in the legend.
• Figure 3: Success rate vs. distance (top panels) and metallicity (bottom panels), using a uniform prior (left panels) and a unimodal beta distribution (right panels) with a peak corresponding to $77\pm7$% (based on the performance of the classifier during development). We notice a small decrease in the success rate with distance and a relatively flat behavior with metallicity, especially in the case where a prior is implemented (see Sect. \ref{['s:discuss_literature']} for details). The number of available classified sources from the literature is indicated by the size of the points and the colorbar on the right.
• Figure 4: Fraction of predicted population with metallicity per class. Despite the presence of large errors and important physical and observational biases, there are noticeable trends of each population with metallicity (see Sect. \ref{['s:populations_with_metallicity']} for details). The symbol size and color for each galaxy reflects the corresponding sample size. For clarity we assign each galaxy to an integer ID, shown in the legend at the top.
• Figure 5: Luminosity functions for all sources identified as RSGs in M31, M33, and NGC 6822. We note the presence of some very luminous sources for M31 and M33 with $\textrm{log}(L/L_\odot)\,$>5.5, indicated by a dashed line (see Sect. \ref{['s:luminosity_functions_RSGs']}).