This morning, as I was walking from Tekniska Högskolan metro station to AlbaNova through the Siberian cold which has hit Stockholm, I was thinking about even colder temperatures than the -15 C that I felt on my skin. Did you know that the temperature of Universe was only 3 Kelvin (-270 C) when the first stars were born?! At least that’s what the authors of an article published in ’Nature’ this week claim to have measured and until proven differently they might well be right… Having worked on the Cosmic Dawn (a popular name for the era during which the first stars were born) for many years I was baffled by this news because this temperature is much lower than we thought it was. How could it have been so much colder?

The results in the article were obtained by an experiment called EDGES (Experiment to Detect the Global Epoch of Reionization Signature). The American authors of the paper have over the past decade been trying to measure a signal from hydrogen atoms in the young Universe. When the first stars formed they caused the neutral hydrogen in the Universe to produce this signal. Its strength depends on the temperature of the hydrogen gas and it was produced with a wavelength of 21 cm. However, by the time it reaches us the wavelength has increased to between 1.5 and 5 m (corresponding to frequencies 200 – 60 MHz) since the radiation travelled to us through an expanding Universe. The particular measurement reported in the paper is at a frequency of 78 Mhz, which corresponds to a wavelength of almost 4 m. This means the signal was produced 13.6 billion years ago when the Universe was only 200 million years old!

What makes EDGES special is that it consists of a single antenna which, at least in its original form, fitted in a suitcase. That’s useful because the team of people who worked on the experiment (led by Judd Bowman) is based in Arizona but do their measurements in the desert of Western Australia.Why choose such a remote location? Obviously, not many people live there so there are almost no radio signals made by humans. At other locations, human radio signals can cause big problems; Just think of the FM radio band which runs from 87 to 108 MHz, right in the middle of interesting frequencies for seeing the effect of the first stars. Not picking up these transmissions helps a lot. However, this still doesn’t make it easy to see the far away 21-cm signal from the time of the first stars. The entire sky is filled with bright radio radiation from relativistic electrons and hot plasmas in our own galaxy which is 100,000s, if not million times, stronger than the signal EDGES is trying to see. No Australian desert can help you to avoid the Milky Way!

But similar to people being able to see a small tree on the slopes of a huge mountain, the radio emission from the early Universe can be separated from that of the Milky Way. When you look at different frequencies, the Milky Way radiation shows the same distribution; in other words, it varies in a very regular way with frequency. The 21-cm signal on the other hand is expected to vary significantly from one frequency to another. This is the key to separating them. But since the contrast is quite big, you have to be very sure that your telescope doesn’t artificially produce small variations due to the electronics, interference signals, the Earth’s ionosphere and so on. To get an accuracy of 1 in a 100,000 you need to put in a lot of hard work! In the case of EDGES, they worked their socks off for no less than 10 years out by regularly testing the performance of the antenna in the desert and by frequently testing all the electronic components in the lab.

So after all this hard work, they went out and did the measurements and found a signal! However, something was wrong since the signal was about twice as strong as anything they had expected. Clearly there must be a problem with the antenna. So they went back, changed some things and tried again, only to obtain more or less the same result. They moved the antenna to a different location and measured again; still no change! So after trying many different things and not being able to get anything except a strong signal and not being able to explain it with anything else they decided that it might be real and coming from the time of the first stars.

However, if it is real, it would mean that the temperature of the Universe at the time this signal was produced was at most 3 Kelvin. However, the absolute lowest temperature expected is about 7 Kelvin. Now the difference might not seem much but the problem is that we know of many processes which can increase the temperature but not really of any which can lower it. So, there is a really a problem here, the first stars woke up to a very, very chilly Cosmic Dawn.

However, if we don’t know of any processes which can reduce the temperature then we could try to come up with some. This is exactly what the author of a second paper in Nature did. What is needed is something to cool the early Universe, what could it be? Modern cosmology relies on most of the matter in the Universe to be dark matter, unobservable except through its gravitational effects on normal matter. Now, the type of dark matter which works best in detailed models is known as “Cold Dark Matter” because it’s, well, cold. What if this dark matter would actually interact a little with normal matter? Then the normal matter would lose some of its energy, its heat, to the cold dark matter. It does not need to interact much, just enough to make the temperature drop from 7 to 3 Kelvin. Rennan Barkana, the author of this second paper, worked out the numbers and found this could work. But only if the cold dark matter particles are not too heavy and interact sufficiently with normal matter.

Is this a reasonable explanation? It’s not a type of dark matter particle which is often considered but since we do not know what dark matter is, it’s hard to rule out. Still, it’s an odd result and perhaps the simplest solution is just that despite their best efforts, the EDGES team missed something and the signal is not at all from the time of the first stars but caused by a subtle effect in their equipment. That’s why everyone, including the EDGES team, is hoping another team with a similar radio antenna will confirm their result. Luckily there are several other experiments active which could try this, SARAS, LEDA and possibly NenuFAR, a spin-off from the LOFAR project. However, all of these experiments also need to do the hard work of understanding the tiniest details of their antennae and electronics and how these interact with the radio signals arriving from space and Earth… It may take a while before they are ready.

In the meanwhile I expect there will be a flurry of papers as a result of EDGES result and the proposed cooling by cold dark matter particles. Perhaps the necessary properties for the dark matter particles are already ruled out by some other effects such particles would cause either in a laboratory here on Earth or out in the Universe? Perhaps there are other ‘exotic’ explanations for the detection of such a strong 21-cm signal? We trying to understand a period in the history of the Universe of which we know very little so there is a quite some room for new ideas. There is a lot to look forward to!