Australasian Science: Australia's authority on science since 1938

Clusters of Colour

Omega Centauri

Prominent examples of globular clusters in the southern sky include Omega Centauri.

By Christopher Usher and Duncan Forbes

Determining why some globular clusters are blue while others are red is at the heart of understanding how galaxies assembled.

Globular clusters are among the most spectacular objects visible in the night sky with a small telescope. Prominent examples in the southern sky include Omega Centauri and 47 Tucanae.

Since they are made of stars of the same age and chemical composition but different masses, studies of globular clusters in our own galaxy have helped us to understand how populations of stars evolve. Now studies of globular clusters in other galaxies are providing important clues about how galaxies form and evolve.

As some of the oldest objects in the universe, globular clusters are fossils of galaxy formation that provide a unique view of how galaxies assembled. Since they are made up of hundreds of thousands, or even millions of stars, globular clusters can be studied at much greater distances than individual stars.

Massive galaxies have many globular clusters, with our own medium-sized galaxy having about 160 and the most massive galaxies at the centres of galaxy clusters having tens of thousands.

Globular clusters are either blue or red, and why there should be two distinct colour subpopulations has been a matter of debate. The two subpopulations are seen in almost every massive galaxy, suggesting that the process causing the two subpopulations has occurred universally.

Since globular clusters become redder as they age, a large age difference between the two subpopulations could explain the observed colour difference. However, previous work has shown that the age difference between red and blue globular clusters is far too small to explain their colour difference.

The chemical composition of a globular cluster also affects its colour. Globular clusters that have fewer heavy elements, such as iron or calcium, have bluer colours than globular clusters with more heavy elements.

Most astronomers believe that colour closely traces the amount of heavy elements, so the blue globular clusters are poor in heavy elements and the red globular clusters are rich in heavy elements. The lack of intermediate colour globular clusters is thus due to a lack of globular clusters with an intermediate amount of heavy elements.

However, others have argued that the relationship between colour and the amount of heavy elements is more complicated. They claim that at when heavy elements are at intermediate abundances, small changes in the amount of heavy elements correspond to large changes in colour. If a galaxy had a single population of globular clusters with a wide range of heavy element abundances, this sort of relationship would cause a gap in colour, creating two distinct colour subpopulations.

Our own galaxy hosts two subpopulations of globular clusters that differ in heavy element abundance by a factor of ten. However, one example does not prove that the majority of galaxies have two subpopulations. While the chemical make-up of globular clusters has been studied in several galaxies, until now not enough globular clusters have been studied in an individual galaxy to tell if there are two distinct heavy element subpopulations or only one.

Our team, which includes astronomers in Australia, Chile and the United States, is studying globular clusters around other galaxies. To identify globular clusters and to measure their colours we obtained images of 11 massive galaxies using the Subaru telescope atop Mauna Kea in Hawaii as well as the Hubble Space Telescope. To measure the amount of heavy elements we used one of the twin Keck telescopes, also located on Mauna Kea. Swinburne University of Technology is the only Australian university with direct access to the Keck telescopes, which are the largest optical telescopes in the world.

We used the DEIMOS spectrograph, which acts like a prism to spread out light by wavelength, to observe thousands of globular clusters. By carefully studying how the brightness of the globular clusters change with wavelength we were able to measure the amount of heavy elements in more than 900 globular clusters. This is by far the largest sample of heavy element abundances in globular clusters ever assembled.

In an article recently published in the Monthly Notices of the Royal Astronomy Society we showed that each of the six galaxies with the most measurements have two heavy element subpopulations while the remaining five galaxies have too few measurements to tell if they have two subpopulations or one.

In each case the heavy element abundance differs between the subpopulations by a factor of ten. Furthermore, the blue globular clusters are poor in heavy elements while the red globular clusters are rich in heavy elements. This work has settled the debate about what causes the distinct colour subpopulations.

The blue and red subpopulations differ in other ways than their chemical composition. The red, heavy element-rich globular clusters are found closer to the centre of their host galaxy than the blue, heavy element-poor globular clusters.

Observations with the Keck telescope enabled us to measure how fast the globular clusters are moving. We found that the heavy element-rich globular clusters orbit their galaxies in the same direction and at the same speed as the other stars in the galaxy while the heavy element-poor globular clusters orbit in random directions.

That there are two globular cluster subpopulations with different properties requires that there be two locations where globular clusters form in each galaxy. Since globular clusters are thought to form during burst of intense star formation, the two subpopulations imply that galaxies formed in two stages.

After the Big Bang, the normal matter in the universe was entirely comprised of hydrogen and helium. Since stars get their energy from fusing lighter elements into heavier ones, when stars die they pollute their environments with heavy elements. Thus the amount of heavy elements increased as the universe aged. Since larger galaxies formed stars relatively faster than smaller galaxies, larger galaxies have a larger fraction of heavy elements.

One possible explanation for the formation of two globular cluster subpopulations is that the globular clusters that are poor in heavy elements formed early before the galaxy had been able to create most of its heavy elements. Something must have occurred to shut down the formation of globular clusters while allowing star formation to continue. Later, after the amount of heavy elements had greatly increased, globular cluster formation somehow restarted, forming the heavy element-rich subpopulation. A potential problem with his theory is identifying a mechanism to turn off and on globular cluster formation.

An alternative explanation is that the heavy element-rich subpopulation formed within the galaxy along with most of the galaxy’s stars, while the heavy element-poor subpopulation formed at the same time in nearby lower mass galaxies. The smaller galaxies later merged with the larger galaxy, building up the outer parts of the galaxy we see today.

As we look deeper at galaxies we are seeing more and more evidence of galaxies growing by merging with smaller galaxies. This neatly explains why the heavy element-rich globular clusters are found closer to the centre of the galaxy and why their orbits more closely match the other stars in the galaxy compared with the heavy element-poor globular clusters.

By obtaining the largest sample of globular cluster chemical abundances yet assembled, our work has settled a debate on what causes the red and blue subpopulations by showing their heavy element abundances. The existence of two subpopulations indicates that galaxy formation is a two-stage process. Galaxies may have formed their globular clusters in two bursts, or globular clusters may have formed in galaxies of different sizes that later merged to form the galaxies we see today.

Although future work is required to tell which method dominated, it is quite likely that both played a role in forming globular clusters and the galaxies that host them.

Christopher Usher is a PhD student and Duncan Forbes a Professor in the Centre for Astrophysics and Supercomputing at Swinburne University of Technology.