Our observations of the galaxy NGC 253 were one step in a long process involving many astronomers around the world and many telescopes, with the Hubble Space Telescope (HST) playing a unique role since its launch in 1990 and repair in 1993. I hope here to briefly explain why we performed these observations and what we learned from them.

- Alan Watson

Observations from the Ground

The motivation for our observations begins with a 1985 paper by Chip Arp and Allan Sandage, two of the most influential astronomers of their generation. (Chip Arp pioneered the study of peculiar galaxies with his "Atlas of Peculiar Galaxies". Allan Sandage has worked for decades to determine some of the fundamental parameters of the universe in which we live, such as its age, rate of expansion, and whether it will live forever).

Their paper presents a detailed study of two bright knots of emission coincident with the dwarf galaxy NGC 1569. (Galaxies are often referred to by their number in the "New General Catalogue" or "NGC" of 1888.) The two knots are unresolved in images taken with telescopes on the Earth and appear no different from single stars in our own Galaxy. The question, then, was whether the knots were simply foreground stars in our own Galaxy seen, by chance, in front of the dwarf galaxy or perhaps something more exciting. By analyzing the spectra of the knots (their brightness at many different wavelengths) Arp and Sandage were able to show that the knots were probably not foreground stars. Instead, they suggested that the knots were massive, compact clusters of young stars in the dwarf galaxy itself.

Observations with the Impaired HST

The next significant developments came with the launch of the HST in 1990. Even with its initially-impaired optical performance, the HST gave significantly better images than could be obtained with ground-based telescopes. O'Connell, Gallagher, & Hunter used the HST to obtain images of Arp and Sandage's knots in NGC 1569. Their images showed that these objects were resolved - they appeared larger than the images of stars - and so could not be foreground stars in our own Galaxy. They had to be stellar clusters in the dwarf galaxy itself; Arp and Sandage's suggestion had been proven correct.

Super Star Clusters

The properties of these clusters are astounding. They contain hundreds of luminous stars and have masses of about half a million times the mass of the Sun. However, all of these stars are packed into a size of only about 5 light years, which is about the distance from the Sun to the its nearest neighboring star. These clusters were much brighter and more massive than other known star clusters, and to signify this distinction they were named "super star clusters."

Two questions immediately attracted astronomers to super star clusters. The first was how such clusters could form, that is, how molecular gas, the raw material for making clusters of stars, could be compressed in such enormous quantities into such a small region. This question remains a mystery. The second was whether these clusters might be related to the globular cluster we see in our own Galaxy and in others. Globular clusters contain very old stars and were probably formed when our Galaxy was very young; by studying them we hope to learn something about the origins of our Galaxy. When they were young, globular clusters may well have been similar to the super star clusters in NGC 1569. The super star clusters may tell us something about the youth and formation of globular clusters, and this in turn might tell us something about the formation of our Galaxy.

Dwarf galaxies such as NGC 1569, where super star clusters were first discovered, are normally quite faint and so it is relatively easy to find bright super clusters in them. The impaired HST gave Jon Holtzman and his colleagues the ability to discover similar clusters in the more difficult environments of the interacting galaxy NGC 1275. After this, many more super star clusters were found in similar dwarf and interacting galaxies. These observations demonstrated that super stars clusters are not limited just to dwarf galaxies. Astronomers started to wonder just how common they might be.

Observations with the Repaired HST

In 1993, the Space Shuttle visited the HST and astronauts performed space walks to install new instruments. These new instruments corrected the optical problems that had degraded images and allowed the HST to perform as we had always hoped it would.

One of the questions we decided to address with the repaired HST was just how common super star clusters are. The impaired HST had shown that they existed in dwarf galaxies and interacting galaxies, but the improved performance of the repaired HST would be required to investigate a yet more challenging environment: the centers of starburst spiral galaxies. These galaxies are differ from normal spiral galaxies (like our Galaxy and the nearby Andromeda galaxy) in that they are forming stars much more rapidly. The brightness and richness of features in the their central regions makes even objects as bright as super star clusters difficult to see.

We chose to observe NGC 253, which at a distance of 10 million light years is the nearest starburst spiral galaxy. We took images of the galaxy only a few months after the HST was repaired. To our delight, the images showed four super stars clusters. Star clusters had now been found in all three locations where intense star formation is taking place: dwarf galaxies, interacting galaxies, and starburst spiral galaxies. We concluded that the important factor was the presence of intense star formation and that the type of galaxy was irrelevant. That is, we expect that the formation of super star clusters is "robust", by which we mean that it does not depend on a special property of the galaxy, and so we expect that they will be formed whenever a galaxy is forming stars at a sufficiently high rate.

Our conclusion has important implications for the two questions mentioned earlier. The robustness of the formation of super star clusters suggests that the mechanism by which they form is similarly robust; this will help constrain which theories can and cannot explain these clusters. Secondly, it suggests that the mechanism by which they are currently forming may also have operated in the past, when our Galaxy was forming globular clusters. This strengthens the ties between super star clusters and globular clusters and is one step further along the road to understanding the origin of our own Galaxy.

A Broader View

Our observations illustrate two important themes in modern astronomy. One is that observations and theories rarely exist in isolation, but more frequently are steps in a long scientific process leading, we hope, towards a better understanding of our Universe. The other is that observations of phenomena in the current-day Universe can often give us insight into the early Universe in which our Galaxy was born.