This article is reproduced with the permission of New Scientist for exclusive use by Nova users.

On the hunt for cosmic fossils
16 May 2007
From issue 2604 of New Scientist magazine, page 44-46
by Stuart Clark

Enlarge Out of the darkness

When it comes to galaxies, astronomers have a lot in common with Victorian fathers. They have a pretty good idea of what is involved in their conception and they know what the end product looks like. But the actual birth itself is a total mystery.

Using optical telescopes, astronomers can peer back through space to a time around 1 billion years after the big bang. Even at this early stage there are galaxies, not as fully grown as the ones today, but recognisable as galaxies nonetheless.

To look back further in time means studying microwaves rather than light, and that's when we lose sight of galaxies altogether. Instead, we see a universal bath of radiation that carries the imprint of the way the universe looked just 300,000 years after the big bang. At that point in cosmic history, there were no galaxies - just rippling undulations in the density of the gas that filled space.

The conclusion is inescapable: galaxies formed somewhere between 300,000 years and 1 billion years after the universe began. However, precisely when and how is still anyone's guess. "It is a very important piece in the puzzle and at the moment it is missing," says Jayaram Chengalur of the Indian National Centre for Radio Astrophysics in Pune.

We can start to fill this gaping hole in our understanding by looking at the effect galaxies had on the universe. Their formation caused one of the most profound watersheds in cosmic history. Astronomers call it the era of reionisation, and it was a catastrophic event in which almost every hydrogen atom in the universe was ripped apart. It happened when the first objects began to appear inside galaxies (see Diagram). As they formed, huge amounts of high-energy radiation were released, and this flooded into space, stripping electrons from their nuclei. The question is, what were these first galactic objects?

One school of thought holds that they were stars, blazing their glory into space. Another thinks they were black holes, sucking glowing gas into oblivion. To find out, astronomers are gearing up to look for the effect these celestial objects have on their surroundings. That means building a new generation of telescopes specifically designed to detect radio waves emitted by hydrogen gas in the early universe before it was reionised.

For the first 300,000 years of its existence the universe was so hot that there were no atoms, only free electrons and nuclei. It was less than a thousandth of its current size and it had a temperature of thousands of kelvin rather than the chilly 3 K above absolute zero of the cosmos today. Only when the observable universe had expanded beyond 300,000 light years across did the temperature fall low enough for nuclei and electrons to combine to form neutral atoms, most of which were hydrogen. As they combined, the particles allowed a burst of radiation to flood the universe - and we now know it as the cosmic microwave background.

After this, a period known as the dark ages set in, so called because there was nothing around emitting light at that time. Hydrogen atoms continued to linger during the first stages of galaxy formation, and only when the first objects formed and began radiating did reionisation begin. Eventually it swept through the entire universe, with only one in every 10,000 atoms escaping destruction.

There is a way to detect this crucial period of cosmic history, or at least to detect the disappearance of hydrogen atoms as they ionise. Hydrogen atoms consist of a lone electron orbiting a proton. Both possess a quantum mechanical property called spin, and the spins of the electron and the proton can be in the same or opposite directions. When an electron flips its spin from the same to the opposite direction, which it can do spontaneously, it releases radiation - radio waves, in fact - with a distinctive wavelength of 21 centimetres.

Slice at a time

As this happened in the early universe, by the time these fossil radio waves reach us, their wavelength will have been stretched by their passage across the expanding universe. The further radiation travels, the longer the wavelength becomes. By scanning a range of wavelengths from 1 to several metres, you can focus on hydrogen at different times in the universe's early history. In effect, you can construct a slice-by-slice scan of the era of reionisation. In fact, 21-cm radiation will tell cosmologists far more about the universe than the cosmic microwave background ever did, filling in many of the details from shortly after the big bang right up to the present day.

The main feature astronomers will be looking for is a dwindling signal. As hydrogen was ripped apart during reionisation, it lost the ability to emit radiation. So astronomers expect to see the 21-cm radiation - so-called even though the wavelengths are much longer when they reach Earth - fade as reionisation takes hold. By charting the speed of this disappearance and its distribution, astronomers hope to determine the nature of the first objects and the way galaxies form.

No one knows for sure what these first objects were like. "I think they were probably stars," says Saleem Zaroubi of the University of Groningen in the Netherlands, "but I wouldn't put a lot of money on it." If they were stars, though, they were not like stars as we think of them today. Made almost exclusively from hydrogen and helium, they must have been monstrously large, with each one between 100 and 1000 times the mass of the sun. These stellar behemoths would have pumped out ultraviolet light, ionising the surrounding hydrogen to form a giant bubble.

On the other hand, the first objects to appear in the early galaxies might have been their central black holes. Around 300,000 years after the big bang, the density of matter in the universe varied from place to place. Some clumps of matter were so big that they may have automatically collapsed into black holes thousands of times more massive than the sun. Other clumps might have reached this condition after accreting more matter. Either way, the resulting galaxy would resemble a scaled-down version of a quasar, a type of powerful active galaxy that dominated the universe 4 to 5 billion years later.

There is a way to tell whether galaxies were first characterised by stars or black holes. Astronomers believe mini quasars would have been powered by hydrogen gas falling into the black hole. As the gas spirals to its doom it is violently heated, reaching millions of kelvin. At such extreme temperatures, the black hole would emit much higher-energy radiation than stars, shining X-rays and gamma rays across the universe.

Fledgling galaxies

These can travel further than light from the first stars, so an ionised bubble created by a mini quasar will be larger than that created by a star. It will differ in another respect, too: the further the X-rays and gamma rays travel, the weaker they become. Hence the bubble edges from mini quasars are not as well defined as those from stars.

Deciding between the two options - monstrous stars or mini quasars with black holes - is where the new radio telescopes come in (see "Tune in to the dark ages"). The image slices they produce of the dark ages should allow astronomers to see bubbles of ionised hydrogen showing up as dark spots that form and spread as the dark ages ebb away. In principle, we should be able to see which came first: stars or black holes.

It is more than just a matter of which came first. The speed at which reionisation happened is important too for both star and galaxy formation. As it ionises, hydrogen gas heats to 10,000 kelvin, making it better able to resist gravitational collapse. So stars would form less efficiently, and small galaxies would fail to take shape altogether. And with fewer stars, there will be less ionising radiation produced and reionisation will slow down.

So the pattern of ionisation will give an idea of whether galaxies were forming simultaneously all over the universe, or in fits and starts in individual pockets. At the moment it is anyone's guess.

How easy it will be to untangle all of this is still a moot point. "I think a lot of the interpretation is going to be dependent on which models are used to analyse the data," says Chengalur.

Eventually, NASA's James Webb Space Telescope will come to the rescue. Planned for launch in 2013 and sensitive to infrared radiation, it will search for the stretched light from the first galactic objects. These objects should be sitting in the bubbles of ionised hydrogen revealed by the 21-cm radio telescopes, especially by the proposed Square Kilometre Array.

This light will tell us immediately whether the early galactic objects were stars or quasars. The ambiguity will be over and we'll have everything we need to finally witness the process of galaxy formation.

That won't happen soon, though. It will be no easy task to get the James Webb telescope and the Square Kilometre Array working together. In the meantime, the search for reionisation will proceed with smaller radio telescopes. It's a monumental undertaking, with the first tentative results expected around 2010. "Twenty-one-centimetre astronomy opens a new window on the universe. Whenever you do that, you expect surprises," says Zaroubi.

The bottom line is that we simply don't know what to expect, but whatever we get, it will tell us something new about the universe. "We are on a fishing trip," says Ue-Li Pen at the University of Toronto in Canada, "but it's a well motivated fishing trip."

Stuart Clark is a science writer based in Hertfordshire, UK.

Tune into the dark ages The ultimate telescope to probe the cosmic dark age is the proposed Square Kilometre Array. Using thousands of individual antennas separated by up to 3000 kilometres, the SKA will be equivalent to a single radio telescope with a gargantuan collecting area of 1 square kilometre. Construction is planned to start in 2011 in either Western Australia or South Africa, and though the telescope will not be operating at full capacity until 2020, the first findings are expected as early as 2014. That's a still long time to wait, so three other teams are racing to develop smaller instruments, any of which could make the breakthrough and reveal the identity of the first celestial objects. One group is using the Giant Metrewave Radio Telescope in Pune district, India. GMRT consists of 30 45-metre-wide radio dishes that are sensitive to radio wavelengths of several metres - exactly the wavelengths that the 21-centimetre emission is expected to have after its long passage across space. Radio days The second is the Low Frequency Array (LOFAR) headed by the astronomical institute ASTRON in Dwingeloo, the Netherlands. LOFAR aims to use up to 25,000 simple radio receivers, distributed across northern Europe and connected via the internet. "We have taken images with LOFAR and proved that it works," says Saleem Zaroubi of the University of Groningen in the Netherlands. Now they have to build the rest of the antennas. Construction is expected to begin in earnest this summer and be largely completed within a year. Then the hard work really starts. "We have a mountain to climb. This is going to be tough data to analyse," says Zaroubi. There are three main problems to overcome. For a start the signal from the dark ages lies buried beneath a welter of radio signals from our own galaxy. Then the signals will suffer from a severe case of twinkling due to Earth's atmosphere. While stars appear to twinkle because of shifting air pockets, it is electrically charged particles in the ionosphere that disrupt radio signals, causing them to literally swim around the sky. Astronomers can correct for this movement by comparing the images with bright objects whose positions are well known. Trouble is, no one yet knows what the sky looks like at metre wavelengths. "We will spend a year just constructing our sky map to get to know what the sky looks like at these wavelengths before we start correcting our data," says Zaroubi. Finally, there is the problem of terrestrial interference. Much of the signal coming from the dark ages will fall into the wavelengths used by ordinary terrestrial radio. It's a double-edged sword. On the plus side, it means that each of LOFAR's thousands of antennas are no more expensive than an FM radio. On the minus side, it means that the signal has to fight human radio as well as galactic interference. It is a battle that LOFAR simply cannot win. LOFAR astronomers will be able to see the final chapters of reionisation but none earlier. Enter the Mileura Widefield Array located on a sheep farm miles from nowhere in Western Australia. MWA has 8000 antennas and its remote location means it will suffer from less interference, so it will be able to tunnel back further in time than LOFAR to the dark ages.

For the latest from New Scientiist visit www.newscientist.com