Improving Star Cluster Age Estimates in PHANGS-HST Galaxies and the Impact on Cluster Demographics in NGC 628

Abstract

A long-standing problem when deriving the physical properties of stellar populations is the degeneracy between age, reddening, and metallicity. When a single metallicity is used for all star clusters in a galaxy, this degeneracy can result in ”catastrophic” errors for old globular clusters. Typically, approximately 10 – 20 % of all clusters detected in spiral galaxies can have ages that are incorrect by a factor of ten or more. In this paper we present a pilot study for four galaxies (NGC 628, NGC 1433, NGC 1365, and NGC 3351) from the PHANGS-HST survey. We describe methods to correct the age-dating for old globular clusters, by first identifying candidates using their colors, and then reassigning ages and reddening based on a lower metallicity solution. We find that young ‘Interlopers’ can be identified from their Hα flux. CO (2-1) intensity or the presence of dust can also be used, but our tests show that they do not work as well. Improvements in the success fraction are possible at the ≈15 % level (reducing the fraction of catastrophic age-estimates from between 13 - 21 % to 3 - 8 %). A large fraction of the incorrectly age-dated globular clusters are systematically given ages around 100 Myr, polluting the younger populations as well. Incorrectly age-dated globular clusters significantly impact the observed cluster age distribution in NGC 628, which affects the physical interpretation of cluster disruption in this galaxy. For NGC 1365, we also demonstrate how to fix a second major age-dating problem, where very dusty young clusters with E(B − V) > 1.5 mag are assigned old, globular-cluster like ages. Finally, we note the discovery of a dense population of ≈300 Myr clusters around the central region of NGC 1365. and discuss how this results naturally from the dynamics in a barred galaxy.

A Tale of Three Dwarfs: No Extreme Cluster Formation in Extreme Star-forming Galaxies

“…For example, a maximum E(B − V ) of 1.5 mag was adopted by the LEGUS (Calzetti et al 2015) and PHANGS (e.g., Turner et al 2021;Lee et al 2022) cluster surveys, and by Adamo et al (2010Adamo et al ( , 2011a for the clusters in ESO 185-IG13, ESO 338-IG04, and Haro 11. However, adopting a single, large value like 1.5 mag for the maximum E(B − V ) can result in serious problems in the estimated ages for clusters older than ≈10 Myr, as explored and discussed in detail by Whitmore et al (2020Whitmore et al ( , 2023.…”

“…We believe there are two main reasons for the differences described above: (1) we include Hα photometry directly in the SED fits, and (2) Adamo et al (2010, 2011a) adopted a significantly higher maximum E(B − V ) of 1.5 mag in their age-dating procedure, compared with the 0.1 mag we adopt for nearly all clusters. SED fits that only include broadband colors struggle to break the age-reddening degeneracy and can result in a number of systematic, 'catastrophic' failures in age-dating, with many older clusters with little reddening erroneously fit by a combination of young age and high reddening (e.g., Whitmore et al 2020Whitmore et al , 2023M. Floyd et al, in preparation).…”

“…The main challenge is to break the age-reddening degeneracy, since clusters can appear red because they are older or because they are reddened due to the presence of dust (e.g., Adamo et al 2010;Calzetti et al 2015;Turner et al 2021). Usually there are two or three combinations of age and reddening that match the measured broadband colors of a cluster (e.g., see Figure 4 in Whitmore et al 2023). Including information on whether or not line emission is present, from Hα photometry for example, significantly improves age estimates by distinguishing between very young, Hα-emitting clusters and older ones that no longer show line emission (e.g., Chandar et al 2010b;Fouesneau et al 2012;Ashworth et al 2017;Whitmore et al 2020).…”

“…A high maximum allowed E(B − V ) in age-dating can lead to catastrophic age and mass determinations for some clusters when no narrowband photometry is included (as described in detail by Whitmore et al 2020Whitmore et al , 2023. During the fitting procedure, the same range of reddening (from 0.0 up to some maximum E(B − V )) is typically adopted for all clusters, regardless of how much dust is nearby.…”

The X-Ray Binary-star Cluster Connection in Late-type Galaxies

“…While the PHANGS-HST study estimates ages correctly for many clusters, incorrect determinations have been found to disproportionately affect the ancient GC populations. Whitmore et al (2023) andM. Floyd et al (2023, in preparation) have found that ∼80% of ancient GCs, which comprise ≈5%-10% of the PHANGS cluster catalogs, have been incorrectly age-dated, with both too young ages and too high reddening.…”