WAY out in the spherical halo of the Milky Way lie some 200 mysterious
regions where stars congregate more densely than anywhere else in the Galaxy.
Throughout most of the Galaxy, several light years separate each star from its
nearest neighbour. But in these densely populated areas of the halo, more than
100 000 stars are squeezed into a region just 30 light years across. These dense
knots of stars are called globular clusters, and astronomers have long been
mystified about how they formed.
The stars that make up the clusters are as old as any in the Milky Way. This
means that the clusters must have been created just as the Galaxy itself was
forming. By learning more about how globular clusters formed, astronomers hope
to improve their understanding of what happened during the birth throes of the
Milky Way itself.
Until recently, they had little more than inference and guesswork to go by.
The globular clusters in the Milky Way formed billions of years ago, and there
were no globulars in the process of forming for astronomers to study. But now
the sharp eyes of the Hubble Space Telescope (HST) have picked out dense knots
of extremely young blue stars in other galaxies鈥攎ost of them quite
different from our own. Could these blue clusters be the forerunners of dense
crowds of mature stars? And could they throw light on the origins of our own
globular clusters?
Advertisement
The existence of young clusters in other galaxies was hinted at for some
years before the latest discoveries. In the 1970s, Ken Freeman, then at the
Australian National University, suggested that compact clusters of stars
observed in the Large Magellanic Cloud, one of Milky Way鈥檚 satellite galaxies,
might be young globular clusters. A decade later, astronomers reported seeing
similar objects in several other galaxies. In 1990, Dieter Lutz of the Max
Planck Institute for Astronomy in Heidelberg, Germany, discovered compact
clusters of stars in images of a galaxy called NGC 3597. Unfortunately,
ground-based observations like this, which were made with a 2.2-metre telescope
in Chile, could not resolve sufficiently fine detail for astronomers to decide
whether they were compact enough to evolve into something like a globular
cluster. For gravity to keep the stars bound together for billions of years,
each cluster would have to be no more than 30 light years across. But the
measurements could only pin them down to less than 300 light years.
Enter the HST. Perched above the distorting effects of the Earth鈥檚 turbulent
atmosphere, it can measure the sizes of the young star clusters with
unprecedented accuracy. In 1991, Jon Holtzman of New Mexico State University and
his colleagues used it to investigate blue star clusters in a peculiar merging
galaxy called NGC 1275. Sure enough, the clusters turned out to be less than 30
light years across鈥攆ar too compact to be one of the open weakly bound
clusters that are scattered around galaxies.
Spurred on by these findings, astronomers pointed the HST at several more
galaxies. When Bradley Whitmore of the Space Telescope Science Institute in
Baltimore and his colleagues took pictures of a merging galaxy called NGC 7252,
they discovered some 40 blue point-like objects, with the right brightness and
colour to be young globular clusters, that have formed within the past billion
years. These properties make them 鈥渏ust perfect鈥� as candidates for newly formed
clusters, the researchers say. In the pair of colliding galaxies NGC 4038/4039,
nicknamed the Antennae, Whitmore and Fran莽ois Schweizer of the Carnegie
Institution of Washington found over 700 compact young clusters.
Theorists are now struggling to figure out how and why these young clusters
form where they do. One clue is that many of them are found in galaxies that
bear the telltale signs of recent mergers鈥攆or example they may have
peculiar shapes or streaming tails of gas formed as one galaxy tugs on the gas
in the other. Even before the HST spotted the new young clusters, Keith Ashman
of the Space Telescope Science Institute and Stephen Zepf, then at Johns Hopkins
University, also in Baltimore, had already predicted that new globular clusters
should be born when elliptical galaxies are formed from the merging of spiral
galaxies like our own. This, they said, would explain why elliptical galaxies
tend to have a disproportionately large number of globular clusters compared to
spirals.
But this cannot be the whole story. New globular clusters have also shown up
in a class of galaxies called 鈥渟tarburst鈥� galaxies, which have recently
experienced massive episodes of intense star formation but show no signs of
recent mergers. In August this year, Alan Watson of New Mexico State University
and his colleagues reported the discovery of four compact young clusters in the
central regions of the nearby starburst galaxy NGC 253.
鈥淲e used to think that there was something special about interacting galaxies
that produced these dense star clusters,鈥� says Jay Gallagher and his colleagues
at the University of Wisconsin at Madison, have discovered compact blue clusters
in a number of starburst galaxies. Now Gallagher believes that clusters are
produced as a normal part of the intense star formation that occurs when
galaxies meet. Whitmore agrees: 鈥淲henever you have vigorous star formation, you
seem to get dense young clusters,鈥� he says. 鈥淪ince galaxy mergers are the best
way to trigger a starburst, you would expect to find lots of compact clusters in
those systems.鈥�
But that alone does not explain why the clusters are so dense. One further
clue comes from the regions where they are formed: environments in which there
is plenty of gas under high pressure. Prime locations seem to include gas clouds
shocked by an interaction between two galaxies, dense regions close to the
galactic centre, and 鈥渟tarburst rings鈥� where the gravitational pull of a nearby
galaxy induces waves of high-pressure gas like ripples on a pond. By contrast,
ordinary open clusters鈥攍ike the Pleiades in the Milky Way鈥攁re born
after a collapsing gas cloud starts to form the clumps. Radiation from the
newborn stars then blows the remaining gas away, and the cloud stops collapsing.
For the compact clusters, the precursor clouds must have been much more massive,
and they must have collapsed to an unusually dense state before the stars
formed. 鈥淧erhaps if you鈥檙e under high pressure, you can鈥檛 reach an equilibrium
until you鈥檝e collapsed to an incredibly dense state,鈥� says Gallagher.
This could well be what happened during the early history of the Milky Way,
when the globular clusters we now see were formed. One possibility is that dense
regions of high-pressure gas occurred when massive gas clouds fragmented and
collapsed in the Galaxy鈥檚 turbulent birth. Or perhaps a nearby galaxy, now long
gone, was the trigger.
None of this, of course, means that we can be sure that the young clusters in
distant galaxies revealed by the HST really are forerunners of dense groups of
mature stars like the globular clusters. They are certainly small enough, and
many have just the right brightness and colour. But though the brightness
reveals something about their mass, it does not pin it down precisely. So are
these newly formed clusters massive enough to stay bound for the 10 billion
years that will have to pass before they will look like the old globulars in the
Milky Way?
One clever way to tell is by measuring the velocities of the stars in the
cluster. You can do this by measuring the spectra, and seeing how much the lines
are broadened by the Doppler effect, as some stars move towards the Earth and
others move away. Then, assuming there is just enough gravity in the cluster to
balance this motion and keep the stars from flying apart, you can estimate the
total mass.
Luis Ho of the Harvard-Smithsonian Center for Astrophysics in Massachusetts
and Alexei Filippenko of the University of California at Berkeley have done just
that using the 10-metre Keck telescope in Hawaii. In two forthcoming papers,
they report mass measurements for two of the brightest 鈥渟uper star clusters鈥� in
irregular dwarf galaxies NGC 1569 and NGC 1705. The first cluster has a mass 300
000 times that of the Sun, and the second 80 000 times the Sun鈥檚. This makes
them much too massive for their size to be open clusters, but puts them within
the range for globular clusters in our Galaxy. 鈥淥ur results strongly suggest
that, at least in two cases, we are witnessing globular clusters being formed,鈥�
says Ho.
Of course, that may not be true for all young compact clusters. The pair that
Ho and Filippenko observed are among the most luminous known, and thus among the
most massive. The fainter ones may not contain enough mass to become globular
clusters. 鈥淰ery high-mass objects are rare, so measuring masses of a few bright
ones is not the best way to settle the question,鈥� says Sidney van den Bergh of
the Dominion Astrophysical Observatory in Canada. Holtzman and others defend
this approach, however, arguing that it is at least a step in the right
direction.
Van den Bergh has other worries too. He points out that the distribution of
luminosities, and thus masses, of globular clusters in our Galaxy is very
different from that of young blue clusters seen in starburst galaxies. There are
very few faint, low-mass globular clusters but lots of faint鈥攁nd
presumably low-mass鈥攜oung clusters in starbursts.
But this may not be as much of a problem as it at first seems. Many
researchers point out that the low-mass blue clusters in starburst galaxies may
not live to an old age because they are eventually torn apart by gravity from
the rest of the parent galaxy. Meanwhile, their more massive cousins would
survive to become old globular clusters like the those in the Milky Way. 鈥淚t鈥檚
quite possible that the faint ones are destroyed,鈥� Whitmore says. 鈥淏ut whether
that really happens is still an open question.鈥� Zepf, now at the University of
California at Berkeley, and Ashman are hoping to find the answer by observing
clusters in many different galaxies that are at different stages of a merger. If
the fainter clusters are destroyed over time, they should be commonest in the
galaxies that have undergone a merger most recently. 鈥淭he burden of proof is on
those of us who believe that the mass distribution of these clusters evolves
with time, and we haven鈥檛 quite shown that yet,鈥� says Zepf.
The story that will eventually emerge from studies of the new young clusters
may well go further than detailing the origins of our own globulars and reveal
some of the secrets of the way new stars form as galaxies evolve. Because the
clusters are so dense, and the star formation inside them is so intensive, they
act as extremely efficient stellar nurseries. Around 80 per cent of the gas they
contain is eventually converted into stars, compared to only 5 per cent in less
dense parts of a galaxy. What鈥檚 more, if Zepf and his colleagues are right,
fainter young clusters may dissolve away and spew out their stars into the rest
of the galaxy. If so, the new clusters that the HST is discovering could hold
the secret not only of our own Galaxy鈥檚 ancient globular clusters but also of
the origins of many of the ordinary stars that populate galaxies throughout the
Universe.