Next: Gravitational Lenses Up: Overview Previous: Globular Cluster Dynamics

Young Massive Star Clusters: the search for black holes leads to a number of surprises

A potentially enlightening first step in the study of a million-body system that is many billions of years old would be the study of such a system when it is relatively young, from millions to hundreds of millions of years old. Unfortunately, our own Galaxy does not contain such objects. However, the nearby Large Magellanic Cloud does, and our Galaxy contains a few young and very massive open clusters. (An open cluster may contain up to many tens to hundreds of thousands of stars, with stellar densities in their cores rivaling those of globular clusters. However, an open cluster, unlike a globular, does not have a deep enough gravitational potential well to retain its stars for more than a few hundred million years.)

One particularly intriguing reason to study these young massive clusters is to search for black holes. None of the roughly 150 globular clusters in our Galaxy2 is known to contain a black hole (although there are some controversial claims (26,109)), and only recently was the first black hole found in an extragalactic globular cluster (63). This paucity is understood theoretically. A basic trend in the evolution of a star cluster is that the heaviest objects will sink to the cluster core where they will form a sub-core and interact with each other (103). They will tend to pair off in binaries with each other and become more and more tightly bound through super-elastic encounters with other cluster members (92). Through these encounters, they will kick each other out, and eventually all or almost all of them will be ejected from the core (90).

In the younger stages (a few million to a few hundred million years) of a star cluster's life, this process of sinking to the core and ejecting each other is thought to occur with the cluster's massive stars and the black holes formed from the cluster's most massive stars. At an age of about a billion years, all of the black holes (except perhaps one) will have been ejected, and the process repeats with the neutron stars (as is currently happening with the Galactic globular clusters).

According to this scenario, clusters younger than about a hundred million years should be rich in black hole binaries (which includes both black hole + normal star binaries and pairs of black holes) and colliding wind binaries (which consist of a pair of young massive stars with strong stellar winds), many of which are readily observable in X-rays. The very youngest clusters, less than a few million years old, should be rich in colliding wind binaries but not black hole binaries (they are not yet old enough to have black holes; e.g., the main sequence lifetime of a 25 Msun is roughly 8 million years).

I began an investigation with Chandra of some young massive clusters in the Large Magellanic Cloud with Simon Portegies Zwart and Walter Lewin to gain an understanding of their source populations and dynamics. Our initial study dealt with a 2 - 4 million year old massive cluster in the 30 Doradus region. As expected, we found many colliding wind X-ray sources, but no black hole systems (93). Our next observations were of the binary clusters NGC 1850A and NGC 1850B. Based on their ages and masses, we expected to find many colliding wind binaries in both and many black hole binaries in NGC 1850A. In fact, based on the calculations of Portegies Zwart, NGC 1850A is one of the best young clusters in which to look for black hole binaries. Surprisingly, we found no X-ray sources in either cluster. It is unclear yet whether this is a fluke or whether our thinking about the young stages of globular cluster evolution needs to be revised. Simon Portegies Zwart is running new numerical simulations of binary clusters to investigate this problem on the theoretical side, and we have observations of another pair of binary clusters (NGCs 2136 and 2137, both of which have a number X-ray sources) which will shed some observational light on this problem. A graduate student at the University of Amsterdam will be working on these data to search the archival HST images for counterparts to the X-ray sources for classification.

I am also part of another collaboration which observed the very massive Galactic open cluster Westerlund 1 with much the same aims as above. The most exciting result thus far is our discovery of a slowly rotating pulsar (the spin period is 10.6 sec) associated with this cluster (71), which is only about 4 million years old. This cluster is known to contain stars more massive than 40 Msun (12), which implies that the progenitor of this neutron star was even more massive. Conventional wisdom has long held that the outcome of a supernova explosion -- either a neutron star or a black hole -- was a relatively simple function of the mass of the progenitor, with the dividing line around 25 Msun. It now seems that conventional wisdom is wrong.

Our monitoring of this object has resulted in a very rich data set. We observed the cluster on 2006-Sep-16 as part of the XMM-Newton guest observer program. Four days later, the Swift Burst Alert Telescope detected a 20 millisecond burst from the direction of cluster, which prompted an XMM Director's Discretionary Time observation 1.5 days later. The outburst was seen to come from the slowly rotating pulsar, which placed it in the class of “anomalous X-ray pulsars” which are young pulsars with magnetic fields on the order of 1013 - 1015 G. Our analysis of the change in X-ray properties before and after the burst the outburst resulted in a drastic change in the pulse shape -- from single peaked before the outburst to triple peaked after the outburst. This could happen if the outburst were caused by internal changes in the neutron star which deformed the crust as they moved outward and resulted in multiple hotspots.

Thus far, these young massive clusters have held some very interesting surprises, and there other compelling reasons to continue their study. For example, many “ultraluminous” X-ray sources (those with X-ray luminosities above the Eddington limit of a 10 Msun black hole) have been claimed to be powered by black holes of hundreds to thousands of solar masses (67). Many believe that these black holes could be formed somehow in young massive star clusters (91). The lack of observational evidence of any sort of black hole in a young massive cluster makes this idea a little difficult to accept, and the discovery of even one such system could lend some credence to it.