Authors: N. Bastian and F. Niederhofer
First Author’s Institution: Astrophysics Research Institute, Liverpool John Moores University
Status: Accepted by MNRAS
Can star clusters be host to multiple events of star formation? Data from Hubble Space Telescope seems to suggest that this is the case. While we often assume that all the stars in a stellar cluster have the same age, astronomers have recently found that a majority of intermediate age clusters in the Large and Small Magellanic Clouds display evidence of a spread in the age of their constituent stars. To see where this comes from, let’s take a step back to look at some properties of stars.
Stars aren’t born with just any luminosity or color. If you were to plot stellar color against luminosity (what is known as a Hertzsprung-Russell Diagram or a color-magnitude diagram), you’d find a band where the newly-formed stars tend to lie, known as the main sequence. As the stars progress through their lives, they eventually evolve away from the main sequence, with the most massive stars spending the least amount of time on the main sequence and the least massive stars spending the most time on the main sequence. The point where the star leaves the main sequence is known as the main sequence turnoff. While it’s very difficult to measure the age of any single star, we can often use the main sequence turnoff to get an estimate of a stellar cluster‘s age.
How does this work? In general, stars are also not formed in isolation, something that’s been mentioned previously on astrobites. They are usually formed in groups, from massive clouds of molecular gas often hundreds of solar masses in size. As a result, we tend to assume that all the stars in a stellar cluster are roughly the same age (what’s known as a simple stellar population or SSP) and distance from us. We can then estimate the age of the stellar cluster by looking at the age of the stars at the main sequence turnoff; if all the stars in a cluster are about the same age, then the stars leaving the main sequence should also be about the same age, producing a narrow main sequence turnoff. This age would then be the age of the stellar cluster.
On the other hand, if a stellar cluster contains stars formed from several different star formation events, there would be a spread in the age of stars at the main sequence turnoff, known as an extended main sequence turnoff, or eMSTO. The presence of eMSTOs in many intermediate age clusters has caused astronomers to suggest that these clusters have been host to many star formation events. This problem was previously discussed in this astrobite, which focused on the star cluster NGC 1651.
However, when the authors of the previous astrobite paper investigated NGC 1651, they found that other features indicative of multiple generations of star formation–like a wide sub-giant branch–were not present. This led them to conclude that the extended main sequence turnoff might not be an indication of a spread in age after all. But is this true for other intermediate-age clusters?
NGC 1806 & NGC 1846
The authors of today’s paper have looked at the intermediate-age stellar clusters NGC 1806 and NGC 1846, both located in the Large Magellanic Cloud, to see if they can find an age dispersion in other parts of the color-magnitude diagram. In particular, they focus on the width of the sub-giant branch and the red clump. These clusters are estimated to have age spreads of about 200-600 Myr largely based on their MSTO regions.
They began by making color-magnitude diagrams of the stars in both clusters. The one for NGC 1806 is shown in Figure 1. The blue lines running through the diagram are isochrones, curves that represent the location of stars with the same age. The sub-giant branch stars, in red, are clustered around one isochrone, which seem to indicate that there isn’t a spread in their ages.
The authors then created two synthetic color-magnitude diagrams, one showing the distribution if the eMSTO was caused by an age dispersion, and the other for a simple stellar population with an age of 1.44 Gyr (their best fit to the sub-giant branch). To find the difference between their observations and simulations, they subtracted the expected difference in magnitude for stars in the observed sub-giant branch 1.44 Gyr isochrone. This is shown as the black histogram in Figure 2. After estimating the observational errors, they convolve these with a distribution of stars coming from a simple stellar population (in red). Finally, these are compared with the distribution expected from stars that formed over an extended star formation history (in blue). The 1.44 Gyr isochrone is able to reproduce the peak of the histogram, but fails to catch the tail ends, something the authors acknowledge is therefore unlikely to be caused by errors in our photometry (our measurement of the flux). On the other hand, the model that used a stellar population with an age spread fails to account for the peak of the histogram, suggesting an inconsistency between the sub-giant branch and a large age spread. When they analyze the red clump, they also find the stars are clustered around one isochrone.
The authors repeat their analysis for NGC 1846 and obtain similar results for the other cluster, causing them to conclude that eMSTOs are not caused by an age spread.
Their results are consistent with the assumption that stellar clusters do not exhibit a range of stellar age and are also consistent with a number of findings in the literature that support a lack of age spreads in stellar clusters. Instead the authors point towards stellar rotation or interacting binaries as two possible causes of the eMSTOs. Stellar rotation can change the structure of the star and its inclination angle to an observer, causing it to have a different effective temperature and therefore, color. However, we still don’t fully understand the effects of rotation on the sub-giant branch of the color-magnitude diagram, so it is still possible that this explanation, while consistent with the eMSTO region, could be at conflict with other parts of the color-magnitude diagram. Another possible explanation they offer is the presence of unresolved binaries; if we have unresolved binary stars, then we would be recording the flux of two stars rather than just one. These binaries are not expected to cause significant broadening in either the sub-giant branch or the red clump, but also may not fully explain the eMSTO. The authors also acknowledge and encourage alternative explanations for the eMSTOs as well.
Despite the growing evidence for an alternative explanation to age spreads being the cause of the eMSTOs we observe, it’s probably too soon for us to conclude definitively either way in this debate. At the very least, these results indicate that we still need to to study eMSTOs and their possible causes in greater detail.