Dating the Evaporation of Globular Clusters

Editor’s note: This article, written by AAS Media Fellow Kerry Hensley, was originally published on Astrobites.

Tadpole Galaxy

Figure 1. Tidal forces don’t only have an effect on globular clusters! The dramatic tail of the Tadpole Galaxy (Arp 188) is also the result of a gravitational tug. [NASA]

The most ancient stellar populations in our galaxy are being ripped apart. Globular clusters — massive gravitationally bound collections of hundreds of thousands of stars — have occupied the Milky Way halo for billions of years. Studying globular clusters can help us understand not only how our galaxy formed, but also how it has evolved over the history of the universe. As the Milky Way has evolved, its gravitational potential has changed as well — and the changes in our galaxy’s gravitational pull are recorded in the behavior of globular clusters.

As the stars in globular clusters interact gravitationally, some gain enough kinetic energy to be ejected from the cluster entirely. The shrinking of globular clusters through this process is called evaporation. When the ejected stars escape the gravitational confines of the cluster, the gravitational pull of the Milky Way starts to take over. As the cluster orbits the galactic center, it experiences tidal forces. Much like an unwitting spacefarer approaching a black hole, globular clusters get stretched out by these tidal forces, stringing those escaped stars into a tidal tail or stream (see Figure 1).

We see these tidal tails in many globular clusters, but the question remains: How can we figure out when the tidal disruption began?

Bose, Ginsburg & Loeb 2018 Figure 1

Figure 2. Simulated globular-cluster tidal streams at a redshift of 0. All else being equal, a more compact cluster (bottom row) experiences less tidal disruption than a more extended cluster (top row). [Bose et al. 2018]

Tracing Tidal Evolution

Today’s paper introduces a new technique to estimate the age of tidally disrupted globular cluster streams. While the ages of the globular clusters themselves are usually determined using stellar evolution models, it can be challenging to figure out when the gravitational pull of the Milky Way began to tear them apart.

The authors of today’s paper found a way to pinpoint the time of tidal disruption by considering the evolution of simulated globular clusters orbiting a Milky-Way-like galaxy. They varied the initial position and velocity (with respect to the center of the Milky Way’s gravitational potential) of six otherwise identical globular clusters — they have the same mass (100,000 solar masses) and initial mass function.

The simulated globular clusters were then allowed to evolve forward in time from roughly 13 billion years ago to today. Figure 2 shows the results of the simulations for three of the six clusters, demonstrating the importance of the initial conditions.

Tidal Disruption in 3… 2… 1…

While the simulated globular clusters give us a good sense of what happens during tidal disruption, they aren’t a good representation of what we would actually observe; telescopes aren’t infinitely sensitive, and their magnitude cutoffs can impose some interesting restrictions on observations. To explore this, the authors simulated what the Gaia survey would see when observing these clusters, assuming a sensitivity limit of 20 magnitudes. Figure 3 compares observable and unobservable stars from three of the simulated clusters.

Bose, Ginsburg & Loeb Figure 2

Figure 3. Maps of the sky in galactic coordinates showing three of the simulated clusters. Stars unobservable by Gaia are shown in grey, while observable stars are shown in color, with the color corresponding to each star’s radial velocity. The number of observable stars decreases as the distance to the cluster increases (counterclockwise from top right). [Bose et al. 2018]

Now considering only the stars that Gaia would observe, the authors found that it’s possible to relate the time at which the tidal destruction of the cluster happened to the proper motions and parallaxes of stars in the tidal tails. Specifically, the more scattered the positions of the stars and the less scattered their velocities along the tidal stream, the longer the time elapsed since the cluster was gravitationally disrupted.

Bose, Ginsburg & Loeb Figure 4

Figure 4. Comparison of the disruption time estimated using the positions and proper motions of stars (horizontal axis) and searching backward through the simulations. The proper motion/position method tends to slightly underestimate the time since tidal disruption. [Bose et al. 2018]

To test that their method recovers the correct disruption time, the authors also searched backward in time in their simulations. They were looking for the time at which the motions of the stars in the tidal stream switched from being controlled by the gravitational potential of the cluster to the gravitational potential of the Milky Way — in other words, the point in time at which the cluster was disrupted. Figure 4 shows that the two methods agree reasonably well, but that the magnitude cutoff of their simulated observations skewed the estimate to be smaller than the true value.

This happens because larger stars are brighter and more easily observed, and tend to be found closer to the center of the cluster. As a result, smaller stars are drawn into the tidal stream before larger stars, and the inferred time since disruption tends to be a bit too short. Still, this is a huge step forward along the path to accurately dating the tidal disruption of globular clusters — an exciting prospect for learning more about the Milky Way’s history!


“Dating the Tidal Disruption of Globular Clusters with GAIA Data on Their Stellar Streams,” Sownak Bose et al 2018 ApJL 859 L13. doi:10.3847/2041-8213/aac48c