Globular star clusters: a new wave of astronomy research

Globular star clusters: a new wave of astronomy research

Associate director of CIERA at Northwestern University Shane Larson explores the role of globular star clusters within the Universe.

The modern scientific enterprise is one that is built around mutually supportive tiers of endeavour: experiment, theory and, increasingly, computation. Ideas about natural phenomena and the laws that govern the way Nature works grow organically from the mixing of results and experiments that are carried out in each of these three tiers.

In astronomy, there are very few laboratory experiments that match the conventional image of laboratories and experiments that are conjured in people’s heads when you say the word “science.” The reason is simple: in astronomy, the Universe is our laboratory. The difference between the Universe and a conventional laboratory is that researchers are highly constrained by the fact that the Universe is not a hands-on experiment; everything we discover is based on non-interactive observations. Experiments cannot be controlled, nor can they be repeated. We are confined to our own small corner of the Universe, probing outward with a meagre suite of instruments designed to detect and measure what our senses cannot. As such, our understanding of the Universe is always bound by our current state of knowledge (what we have already discovered) and limited by both the capabilities of our devices and by chance – were we paying attention when a rare event happened? The net result is that over time our understanding of astronomical phenomena dramatically changes and evolves. Our understanding of the nature of globular star clusters is an excellent illustration of this process.

What are globular star clusters?

Globular star clusters are collections of stars, typically hundreds of thousands to a million stars, which remain mutually bound together by their collective gravitational attraction. They exist in the vicinities of large galaxies, orbiting the galaxy in a spherical swarm like a cloud of bees or mosquitoes. The Milky Way has 157 known globular clusters,1 though we expect that there could be as many as 200, with the missing globular clusters obscured from our view on the far side of the Milky Way. Large galaxies, like the massive elliptical galaxy M87, have more than 10,000 globular clusters orbiting them.

The first globular star cluster was catalogued by Ptolemy in 150 AD, known today as Omega Centauri. The name of this cluster belies our early ignorance about their true nature: Omega Centauri was named as if it were a star in the constellation Centaurus! To the unaided eye, it is very stellar in appearance – a bright point of light without any extended structure.

The first inkling that globular clusters were an interesting and unknown astronomical phenomena came decades after the invention of the telescope in 1665, when Abraham Ihle discovered the globular cluster M22 in the constellation Sagittarius. More than a decade later, in 1677, Edmund Halley discovered that Omega Centauri was a similar object, not a star, as had been presumed up to that point.

Like many first-generation scientific instruments, early telescopes lacked the capability to resolve globular star clusters easily. Even through telescopes, globular clusters looked like diffused, fuzzy clouds of light, which in the early days of telescopic astronomy were generically called “nebulae” (“clouds”); globular clusters with their circular appearance were called “round nebulae.” In 1764, Charles Messier was the first person to observe individual stars in a globular cluster, while he was observing M4 in the constellation Scorpius. William Herschel used much larger instruments to observe all the 34 known “round nebulae” at that time and discovered 36 new candidates on his own. He found he could resolve stars in nearly all of them. He coined the name “globular cluster” in 1789, creating the class of astronomical objects we know today.

How has understanding of globular clusters evolved?

Since then, globular star clusters have played a vital role in evolving our understanding of the Universe, beginning with mapping the location of the Sun in the Milky Way,2 to understanding the age of the Universe and the lives of stars. Globular clusters harbour the oldest stars known, so are thus some of the oldest known structures in the Universe.

Their nature makes them excellent astrophysical laboratories, since they are dense stellar environments. A typical globular cluster might have, on order, ten stars per cubic lightyear at its core, compared to the galaxy which has only about 0.004 stars per cubic lightyear in the region of the galaxy around the Sun. With stars existing in such close proximity to each other, there is ample opportunity for them to gravitationally interact, especially over the long history of a globular cluster lifetime.

The dominant mechanism that affects the long-term evolution and structure of a globular cluster is gravity. The globular cluster maintains its identity as an individual object, as well as its overall spherical shape, due to the mutual inward gravitational attraction of all the stars. It is clear, however, that if left to its own devices, gravity should cause the cluster to ultimately collapse on itself. Since globular star clusters exist in the Universe, something forestalls such a catastrophe, supporting the cluster against collapse. Understanding the support mechanism and structural evolution of a globular cluster has been the subject of intense theoretical and numerical simulation for decades.

These investigations have shown that a major player in the support of globular clusters are binary star systems. The basic premise is that binary stars are a source of energy that can be tapped to puff up a globular cluster when it contracts under the influence of gravity. In the dense stellar environment, a single star can have a close encounter with a binary. The result of that encounter is the star picking up energy and throwing it onto a larger orbit. The energy it picks up comes from the binary orbit, which shrinks as a result of losing energy to the interloping star. The star, now on a larger orbit, is part of the population of stars that define the overall size and shape of the outer regions of the globular cluster. This mechanism is a stellar version of the familiar “gravity assist” used by spaceflight engineers to send spacecraft into the outer solar system, whereby a spacecraft gains energy by sucking energy from the orbit of the planet it passes.

Investigating the mysteries of the Universe

What ultimately happens to the binary that was left behind? Such questions cannot be answered with real-time observations, since the processes described here occur on timescales vastly larger than a human lifetime. As a result, astrophysicists have turned to computational modelling, simulating the life and times of a globular cluster numerically. The basic premise is simple: create a nascent globular cluster at the start of its life and evolve both the physical properties of the stars forward in time, according to our best understanding of stellar evolution, whilst also modelling the gravitational interaction of all the stars in the globular cluster with every other star. There are many challenges facing such an endeavour, ranging from our developing understanding of how stars evolve, particularly when they are interacting in close quarters, to the sheer scale of tracking the properties, positions and motions of half a million stars or more. Numerical computations on this scale not only require massive amounts of memory, but also long runtimes.

One way to address these computational challenges is to utilise statistical methods, such as Northwestern University’s Cluster Monte-carlo Code (CMC).3 CMC allows the rapid modelling of fully evolved globular star clusters with a variety of properties that can be matched to the observational constraints of known globular cluster systems. CMC includes the physics relevant to the formation and evolution of binaries in dense stellar systems and, as a consequence, CMC is a valuable tool for predicting what the binary population in globular clusters is today.

Of particular interest are binary systems comprised of stellar remnants in all forms, including:

  • White dwarfs;
  • Neutron stars; or
  • Black holes.

These dead star systems are of intense interest for a variety of reasons. First, they encode information about the population of stars they were born from. Second, stellar remnants are tiny compared to ordinary stars, so when they form binaries they can get extremely close together before they touch and merge. This means that through the many encounters they may experience in the dense stellar environment, they can be driven closer and closer together in a process astronomers call “hardening”, providing the energy to support the structure of the globular cluster. Lastly, tight binary systems made of stellar remnants are excellent sources of gravitational waves.

One of the interesting astrophysical outcomes of simulating globular clusters is that a fraction of the systems containing black holes are ejected, breaking free of the globular cluster and destined to roam in the space around the Milky Way (an area called the ‘halo’). Meanwhile, significant numbers are still retained and have a strong influence on the evolution of the cluster over time. Heavy objects in the cluster, like systems that contain black holes, tend to separate themselves from the other stars in the cluster in a process known as ‘mass segregation’, where the heavier systems migrate toward the centre of the cluster. Since systems with black holes are preferentially heavier, they tend to sink toward the centre of the cluster and congregate there. In such dense environments, they can become members of binary systems and become strong sources of gravitational waves.

A new wave of observatories

Gravitational wave astronomy has catapulted into the public eye with a series of prominent detections by the ground-based gravitational wave observatories – LIGO (in the United States) and Virgo (in Europe). These observatories are sensitive to the penultimate inspiral and eventual merger of stellar remnant binaries comprised of neutron stars or black holes. During these events, the binaries circle each other tens to hundreds of times per second. At earlier times in their lives, stellar remnant binaries circle each other on larger orbits, moving more slowly – only once every 15 minutes or so. During this phase of their lives, they also emit gravitational waves, but the waves are different than those seen by LIGO and Virgo; in this phase the waves are long wavelength or low frequency. These gravitational waves can also be observed, but it requires a much larger observatory – a million times larger than LIGO or Virgo. ESA, with NASA as a junior partner, is planning an observatory called the Laser Interferometer Space Antenna (LISA). Comprised of a constellation of three spacecraft exchanging laser signals over 2.5 million kilometres, LISA will orbit the Sun, trailing behind the Earth in its orbit. LISA’s enormous size makes it sensitive to a wide range of astrophysical sources, including stellar remnant binaries.

The Milky Way itself will be a dense field of stellar remnant binaries; LISA will be able to pick out tens of thousands of individual binaries above a steady background hum of around ten million others. But what about the binaries in the globular star clusters? Our simulations with CMC suggest LISA will be sensitive to around 50 detectable sources from the Milky Way globular clusters,4 with the possibility of several others from much farther away – from the globular clusters in the Andromeda Galaxy two million lightyears away, to the globular clusters around galaxies in the Virgo Cluster, which are 65 million lightyears away.

Despite a seemingly small number of LISA detectable binaries in globular star clusters, compared to thousands in the Milky Way, the globular cluster population is unique and important. The globular cluster binaries from dynamical interactions between stars in the dense stellar environment and will have different characteristics than binaries, of which will form in the galaxy. One of the notable differences is the elliptical shape of the orbits – eccentricity – that results from dynamical formation. Eccentricity is a property that can be measured in the gravitational wave data; LISA observations of globular cluster binaries will provide a unique opportunity to probe the galaxy’s globular cluster system and compare the resulting observations and detections to the predictions from our best computational models.

In the finest traditions of astronomy, this will confirm our ideas about the evolutionary processes in globular clusters, or it will push us to improve the physics we currently use to describe globular clusters.


  1. Harris 2010; arxiv/1012.3224
  2. Shapley, PASP 1918
  3. Chatterjee et al., MNRAS 2013
  4. Kremer et al. PRL 2018

Shane L. Larson
Research Associate Professor
Associate Director
Northwestern University
+1 (847) 467-4305


Please enter your comment!
Please enter your name here