MOST PEOPLE THINK OF GALAXIES as huge islands of stars, gas and dust that populate the universe in groups and clusters. But theory has predicted that some galaxies, called “dark galaxies,” have few if any stars and are made predominately of dense gas. Because they are devoid of stars, dark galaxies have been impossible to detect directly – until now. An international team of astronomers, using the European Southern Observatory’s 8.2-meter Very Large Telescope (VLT) in Chile, has detected several dark galaxies by observing the fluorescent glow of their hydrogen gas illuminated by the ultraviolet light of a nearby quasar.
The galaxies detected by the team are nearly 11 billion light years away, which means they existed at an early time in the 13.7 billion-year-old Universe. Dark galaxies are thought to be the building blocks of modern-day star-forming galaxies, either through mergers with star-forming galaxies, or by feeding them gas along filaments that connect the Universe’s galaxies in a kind of cosmic web.
Three members of the team that made the detection spoke recently with The Kavli Foundation in a roundtable discussion about how they made the discovery, what it means, what questions are still unanswered, and what they plan next. The participants were:
- Sebastiano Cantalupo – Postdoctoral Fellow at the University of California, Santa Cruz, the Astronomy and Astrophysics Department. He studies the high redshift universe, and in particular the space between galaxies known as the Intergalactic Medium (IGM) to learn about galaxy formation and evolution. Dr. Cantalupo also studies how the first stars and galaxies ionized the fog of neutral hydrogen gas in the early universe, making it transparent to light.
- Martin Haehnelt – Professor of Cosmology and Astrophysics, Institute of Astronomy and the Kavli Institute for Cosmology, University of Cambridge. An observational cosmologist interested in the emergence of structure during the epoch of reionization, he was a member of the initial science working group for the European Extremely Large Telescope project in Chile, anticipated to see First Light in the early 2020s.
- Simon Lilly – Professor of Observational Cosmology, Institute for Astronomy at the Swiss Federal Institute of Technology in Zurich, Switzerland. His group seeks to understand the formation and evolution of galaxies, and uses observational data on the distant universe obtained from ground-based and space-based observatories. Dr. Lilly is involved in instrumentation projects for the VLT and the James Webb Space Telescope.
SEBASTIANO CANTALUPO: Dark galaxies are similar in some respects to those we see today, as they are composed of dark matter and gas; but for some reason they have not been able to form stars. As a result, they cannot be detected with our optical telescopes. Some theoretical models have predicted that dark galaxies were common in the early universe when galaxies had more difficulty forming stars – partly because their density of gas was not sufficient to form stars – and only later did galaxies begin to ignite stars, becoming like the galaxies we see today.
SIMON LILLY: This seems to be an epoch when the Universe as a whole was forming stars at a peak rate – about 20 times faster than today. It also seems to be a key time for the growth of black holes, because we see at that time a peak in the number of bright quasars where these growing black holes reside.
TKF: So this is a time when we had galaxies maturing very quickly with very rapid starbursts, but we also had these dark galaxies – these dense clouds of gas that are not yet forming stars. Should we consider dark galaxies to be precursors to modern-day galaxies?
MARTIN HAEHNELT: To make that same point a little more concrete, we expect the precursor to the Milky Way was a smaller bright galaxy that merged with dark galaxies nearby. They all came together to form our Milky Way that we see today.
TKF: Did dark galaxies exist only early in the history of the universe?
HAEHNELT: They exist today but are not easy to see. There is actually a firm prediction from our current theory of galaxy formation that there should be many dark galaxies in our own local group of galaxies, which includes the Milky Way and Andromeda galaxies but also many smaller objects. Many small satellite galaxies in the Local Group are actually expected to be dark galaxies. However, many of them lack enough stars to be detected by starlight, and they also have very little gas at the present time.
TKF: And are these dark galaxies relics of the very early universe, or have they formed more recently?
HAEHNELT: They’re a mixture. Some should be very old, some would have formed later, and some may have joined the Local Group relatively recently.
"We expect the precursor to the Milky Way was a smaller bright galaxy that merged with dark galaxies nearby. They all came together to form our Milky Way that we see today." – Martin Haehnelt
LILLY: By detecting the emission from hydrogen gas within them. This emission is generated when ultraviolet light shines onto the gas and causes its atoms to excite. When they de-excite, these atoms emit photons with a very particular wavelength that we can detect and recognize. Now, the required ultraviolet light permeates the universe but usually the resulting emission is very, very faint. Our approach was to look for dark galaxies in places where the ambient ultraviolet light would be much brighter than the usual background levels – and that was in the vicinity of bright quasars. In the neighborhood of a quasar, this extra UV light boosts the emission from the gas in dark galaxies to levels that we can in principle detect with powerful telescopes.
TKF: If you're just looking around quasars for these dark galaxies, aren’t you underestimating how many there might be?
LILLY: Yes, but the region of space where we are detecting dark galaxies extends out to something like 10 or more Megaparsecs away from the quasar (more than 30 million light years), so we are not just looking at a very small region around the quasars.
CANTALUPO: At the same time, the success of our survey depends on using one of the brightest sources of light that we know of in the universe, which is why quasars are so important.
LILLY: And, using quasars also means that we may be able to use dark galaxies to learn something about the quasars illuminating them – for instance, how long they have been shining at their current very high level, and whether the light from the quasar is emitted in all directions or is just beamed in some particular directions.
"I wonder if we can indeed use this technique to see the emission of filamentary gas in the cosmic web. ... That has been something of a Holy Grail for many, many years and I think this most recent discovery of dark galaxies is a significant step toward the goal." – Simon Lilly
TKF: How large are the dark galaxies you’ve detected?
CANTALUPO: These things are probably as small or smaller than the Magellanic Clouds near the Milky Way. That means they are probably five or six kilo parsecs, [or about 16,500 to 19,800 light years across. By comparison, the Milky Way Galaxy is about 100,000 light years across.] So they are really dwarf galaxies.
TKF: Which makes it even more remarkable that you can detect them from 11 billion light years away.
CANTALUPO: That’s because they have a lot of gas, and the nearby quasar is illuminating all this gas to make it bright enough to be seen.
TKF: Do you expect dark galaxies are also embedded in these filaments?
TKF: What are the prospects for detecting dark galaxies that are even farther away than the ones you found?
LILLY: A number of instruments are being designed and constructed, which will enable us to take our observations to the next level in sensitivity. Our current observations were done with a narrow filter to isolate the emission that fluorescent gas produces. One can use a spectrograph to split the light up much more finely, and that gives you a big potential increase in sensitivity to isolate this emission. But there’s currently a cost to that approach, because it significantly reduces the area of the sky that you can look at.
However, the next generation of instruments is expected to enable us to look at a large area, about an arc minute to a side, [which is equal to about 1/30th the size of a full moon], and across a wide range of wavelengths and with a high spectral resolution.
The MUSE (Multi Unit Spectroscopic Explorer) spectrograph is one of several competing instruments – another spectrograph will be on the Keck telescope in Hawaii. MUSE should be operating in about a year on the VLT in Chile. With long exposure times on MUSE, we should have the sensitivity to see these filaments in the cosmic web – particularly if UV light from a nearby quasar boosts the gas emission. We are quite optimistic about this, and the question is really whether we can make that detection first.
CANTALUPO: This first survey was sort of a test of the technique. Now that we know it works, we will study the regions around ten or so quasars and look for more dark galaxies. This new survey will be conducted in November, from the Keck telescope on Mauna Kea in Hawaii. But before that, in October at the VLT in Chile, we’ll keep studying the dark galaxies we’ve already detected. In particular, we are going to do spectroscopic analyses of these objects. In the future it will also be very important to get images from space, so the Hubble Space Telescope will be very valuable.
HAEHNELT: So far, we have relatively little direct information about the physical properties of dark galaxies from our observations. Computer models have helped us tremendously to understand what it is that we’re seeing. But with additional observations, we can learn more about the underlying properties of these objects, for example how much dark matter you would expect these objects to have. And depending on what dark matter is actually made of, there are very different predictions for how many of these dark galaxies there should actually be. If we manage to count accurately the numbers of dark galaxies that we see around quasars, then this might actually allow us to get a handle on discriminating between competing theories about what dark matter might be.
"The questions that have motivated us for seven years now are, 'How do galaxies form their stars?' and 'How do we look at the earliest stages of galaxy formation?'... [T]hese are now questions we can begin to address because of this new technique." – Sebastiano Cantalupo
TKF: As each of you moves forward, what are your biggest questions about dark galaxies?
HAEHNELT: I would really like to know the minimum mass of galaxies that are able to efficiently produce stars. By observing a sufficiently large sample of the dark galaxies we have detected, we might be able to answer this question.
LILLY: I am first and foremost an observer, and I wonder if we can indeed use this technique to see the emission of filamentary gas in the cosmic web, and if so, how close are we to seeing that? That has been something of a Holy Grail for many, many years and I think this most recent discovery of dark galaxies is a significant step toward the goal.
CANTALUPO: The questions that have motivated us for seven years now are, “How do galaxies form their stars?” and “How do we look at the earliest stages of galaxy formation?” Today, we can only see what dark galaxies look like, estimate their mass, say a few things about the efficiency at which they form any stars at all, and speculate why they are there and their ultimate fate. But these are now questions we can begin to address because of this new technique.
LILLY: As Sebastiano said, we started this project back in 2004. Eight years later, it's nice to see the little acorns that we planted then now growing into oak trees.