The First Star Clusters

Title: The Luminosity of Population III Star Clusters

Authors: Alexander L. DeSouza and Shantanu Basu

First Author’s Institution: Department of Physics and Astronomy, University of Western Ontario

Status: Accepted by MNRAS

First light

A major goal for the next generation of telescopes, such as the James Webb Space Telescope (JWST) is to study the first stars and galaxies in the universe. But what would they look like? Would JWST be able to see them? Recent studies have suggested that even the most massive specimens of the very first generation of stars, known as Population III stars, may be undetectable with JWST.

But not all hope is lost–one of the reasons why Population III stars are so hard to detect is that, unlike later generations of stars, they are believed to form in isolation. Later generations of stars (called Population I and Population II stars) usually form in clusters, from the fragmentation of large clouds of molecular gas. On the other hand, cosmological simulations have suggested that Population III stars would form from gas collected in dark matter mini-halos of about a million solar masses in size which would have virialized (reached dynamic equilibrium) by redshifts of about 20-50. Molecular hydrogen acts as a coolant in this scenario, allowing the gas to cool enough to condense down into a star. Early simulations showed that gravitational fragmentation would eventually produce one massive fragment–on the order of about a hundred solar masses–per halo.  This molecular hydrogen, however, could easily be destroyed by the UV radiation from the first massive star formed, preventing others from forming from the same parent cloud of gas. While Population III stars in this paradigm are thought to be much more massive than later generations of stars, they would also be isolated from other ancient stars.

However, there is a lot of uncertainty about the masses of these first stars, and recent papers have investigated the possibility that the picture could be more complicated than first thought. The molecular gas in the dark matter mini-halos could experience more fragmentation before it reaches stellar density, which may lead to multiple smaller stars, rather than one large one, forming from the same cloud of gas. These stars could then evolve relatively independently of each other. The authors of today’s paper investigate the idea that Population III stars could have formed in clusters and also study the luminosity of the resulting groups of stars.


Screenshot 2015-03-17 08.03.28

Figure 1 from the paper showing the evolution of a single protostar in time steps of 5 kyr. The leftmost image shows the protostar and its disk at 5 kyr after the formation of the protostar. Some fragments can be seen at radii of 10 AU to several hundred AU. They can then accrete onto the protostar in bursts of accretion. The middle time step shows a quiescent phase. There are no fragments within 300 AU of the disk and no new ones are forming so the disk is relatively smooth–the ones that already exist were formed during an earlier phase raised to higher orbits. The right most image shows the system at 15 kyr from the formation of the protostar, showing how some of the larger fragments can be sheared apart and produce large fluctuations in the luminosity of the protostar as they are accreted.

The authors of today’s paper begin by arguing that the pristine, mostly atomic gas that collects in these early dark matter mini-halos could fragment by the Jeans criterion in a manner similar to the giant molecular clouds that we see today. This fragmentation would produce small clusters of stars that are relatively isolated from each other, so they are able to model each of the members in the cluster independently. They do this by using numerical hydrodynamical simulations in the thin-disk limit.

Their fiducial model is a gas of 300 solar masses, about 0.5 pc in radius, and at a temperature of 300 K. They find that the disk that forms around the protostars (the large fragments of gas that have contracted out of the original cloud of gas) forms relatively quickly, within about 3 kyr of the formation of the protostar. The disk begins to fragment a few hundred years after it forms. These clumps can then accrete onto the protostar in bursts of accretion or get raised to higher orbits.

Most of the time, however, the protostar is in a quiescent phase and is accreting mass relatively smoothly. The luminosity of the overall star cluster increases during the bursts of accretion, and it also increases as new protostars are formed. The increasing luminosity of the stellar cluster can make it more difficult to detect single accretion events. For clusters of a moderate size of about 16 members, these competing effects result in the star cluster spending about 15% of its time at an elevated luminosity, sometimes even a 1000 times the quiescent luminosity. The star clusters can then have luminosities approaching and occasionally exceeding 108 solar luminosities. Population III stars with masses ranging from 100-500 solar masses on the other hand, are likely to have luminosities of about 106 to 107.

These clusters would be some of the most luminous objects at these redshifts and would make a good target for telescopes such as ALMA and JWST. We have few constraints on the star formation rates at such high redshifts, and a lot of uncertainty in what the earliest stars would look like. So should these exist, even if we couldn’t see massive individual population III stars, we may still be able to detect these clusters of smaller stars and gain insight into what star formation looked like at the beginning of our universe.

About Caroline Huang

I'm a 4th year graduate student in astronomy at Johns Hopkins. I work on observational cosmology, specifically getting distances from variable stars. I did my undergrad at Harvard, where I was a joint physics and astrophysics concentrator. I love traveling, cool weather, food, books, and Oxford commas.


  1. What about the conditions of the early universe allowed Population III stars to reach such massive masses/luminosities? You mention that the accretion is happening smoothly. Why does it happen more smoothly than the accretion we see today?

  2. It would be indeed super cool if we could observe stars that were formed in the beginning of the universe! One question that came to my mind is how we can observe the very first stars? Because I remember reading that such objects were not observable because of the rapid expansion that happened right after the Big Bang.

  3. How would molecular hydrogen act as a coolant?

  4. Population III stars puzzle me quite a bit. How could it be isolated if they are to be one of the oldest stars. Wont these stars tend to be a part of the bulge for say a galaxy? Surely stellar material gets absorbed to contaminate the star.

  5. Seems like an interesting paradigm shift of the way we view population III stars. They may not be lone stars after all if they are able to form in clusters like pop I and II stars.

  6. How prevalent do we believe Pop. III stars are compared to Pop. I and II stars? It will be interesting to see if future data from these telescopes can provide more insight into this group of stars.

  7. Following up on Jennifer’s question above, how will we determine if these stars are very difficult to observe or if they are just very few in number?

  8. Do you know how long such PopIII stars existed, basically what the lifetimes of such early stars would have been and how long before there wasn’t enough material left for them? I imagine they would have had fairly short timescales given how much mass is involved.

  9. Are there any known PopIII stars that are “close” to our solar system?

  10. Can observations of the first stars/star clusters in turn constrain cosmology at high redshifts? E.g. can we use the fact that stars existed at some redshift to argue something about when dark matter halos of a certain size virialized?

  11. This is very interesting. Once Alma and JWST begin working on this, will it be immediately obvious whether or not these objects exist or will this be more of a long-term study?

  12. Very interesting post! I am always curious about what the new, very expensive, telescopes are attempting to observe, and consequently, help us prove.



  1. Where are the Sun’s grandparents? | astrobites - […] III stars maintain their true identity and remain metal-free? They have to escape from the cloud in which they formed. There are two…

Leave a Reply