Properties of Stellar Clusters and their Relation to Molecular Gas in the Andromeda Galaxy

Abstract

The apparent age and mass of a stellar cluster can be strongly affected by stochastic sampling of the stellar initial mass function, when inferred from the integrated color of low mass clusters ($\lesssim$ $10^4$ $M_{\odot}$). I use simulated star clusters to show that these effects are minimized when the brightest, rapidly evolving stars in a cluster can be resolved, and the light of the fainter, more numerous unresolved stars can be analyzed separately. I show the success of this technique first using simulated clusters, and then with a stellar cluster in M31. This method represents one way of accounting for the discrete, stochastic sampling of the stellar initial mass function in less massive clusters and can be leveraged in studies of clusters throughout the Local Group and other nearby galaxies. Next I present results for the Panchromatic Hubble Andromeda Treasury (PHAT) cluster sample, which is the largest uniformly derived extragalactic cluster sample to date, containing 2753 clusters. I determine the ages, masses, and extinctions of this cluster sample and demonstrate their accuracy through the use of synthetic clusters. The parameter estimates were done using two methods: CMD fitting of the resolved stars and integrated light fitting using discrete population models. I summarize the most accurate estimates for each cluster, and find that the distributions in age and mass are similar to the cluster distributions in other Local Group Galaxies. Finally, I combine several astronomical data sets to investigate the life cycle of molecular clouds in the Andromeda Galaxy. The primary data sets I use are the PHAT cluster sample and a molecular cloud catalogue that is constructed from new high spatial/spectral resolution (20 pc, 1 km/s) CARMA observations. Several ancillary data sets, including H$\alpha$ and Spitzer IR emission maps are also used, taking advantage of broad wavelength coverage to search for indicators of star formation with different timescales. The distribution of the youngest clusters shows a strong correlation with the molecular cloud distribution, while no correlation is evident for clusters greater than 30 Myr. Each molecular cloud in the sample is classified as a star-forming cloud or a non-star forming cloud, based on the presence of any one of several star formation indicators. About 60\% of the clouds in the sample were found to be associated with massive star formation. Based on the comparison between these observations and the results from a Monte Carlo simulation, I estimate a total lifetime of 13-30 Myr for molecular clouds in M31.