First Light
Switching on Stars at the Dawn of Time


Emma Chapman

This page:


First Light



index pages:

Notes: Glossary of acronyms and technical terms
CMBcosmic microwave background radiation
DCBHdirect collapse black hole
EDGESExperiment to Detect the Global EoR Signature, which produced an unexpected result for which explanations are still uncertain
EoREpoch of Reionisation, a stage in the development of the Universe
Jeans massthe mass (a function of size and temperature) at which a system becomes gravitationally bound
MACHOmassive compact halo object, one proposed constituent of dark matter
MetalsAstrophysicists have their own definition of metals.
Segue 1a dark-matter-heavy dwarf galaxy, relatively near the Milky Way
TARDISTime and Relative Dimension in Space, a vehicle for four-dimensional travel


stars, classified by age (in reverse numerical order)
WIMPweakly interacting massive particle, a likely constituent of dark matter
First Light
Switching on Stars at the Dawn of Time

Copyright © Emma Chapman, 2021

Introduction The Universe began. There were not just stars, but first stars, and second and third stars for that matter. What we experience now is only one stage of a much bigger cosmological lifetime [...]. It’s a lifetime that we can be pretty smug about understanding. We have observations of young and old stars, galaxies that are ancient and galaxies that are newly formed. We live in a time with unprecedented access to the Universe and its history, and our ability to fill in those gaps in knowledge has increased at a lightning-fast pace.
Despite the exponential increase in technology and progress, there is a period in our Universe that, until recently, we had no observations of at all. From 380,000 years after the Big Bang to about 1 billion years after it, the Universe has remained in the Dark Ages. The first stars were born less than 200 million years after the Big Bang.

In human terms, the missing cosmological data is equivalent to missing everything from the moment of conception to the first day of school, perhaps apart from a single ultrasound. It may be a small fraction of time compared to the total lifetime, but when you consider how formative these early years are for humans, it is no wonder that astrophysicists quake at this much missing data when it comes to the history of our Universe. What incorrect conclusions are we coming to about the stars around us or how the Universe is behaving now, because of this lack of data?

Because I am an astrophysicist, in this book I will refer to all chemical elements other than hydrogen or helium as metals.
The most recent generation, Population I stars, are young stars that have lots of metals inside them. They are luminous, hot and live in a galaxy’s disk. Population II stars are older and have fewer metals. They reside in the center of the galaxy or outer halo. It doesn’t take much imagination to carry on down this road and ask: what about the oldest stars? The stars with no metals at all, the stars that started it all. Where are they? The first stars produced the first metals, seeding the Universe and enabling galaxies to form. They were metal-free to start with and we call them Population III stars.*

* The counter-intuitive numerical order of these populations is an artifact of their historical discovery and grouping.


Over the Rainbow
Another example occurs in reindeer, which have eyes adapted to process ultraviolet (UV) light. In the Arctic, visible light is low, but the snow reflects a lot of UV light, allowing reindeer to track predator/prey urine and highlight lichen, their major food source.*

* So if you think about it, with the help of Rudolph, Santa can identify counterfeit money and even trace amounts of cocaine and amphetamines. An uncomfortable new dimension to Santa finding out whether you have been naughty or nice.


Santa Claus

The shapes of stellar spectra are very well approximated by curves called blackbody curves. A ‘blackbody’ is an idealised object that absorbs all incident radiation and emits in a continuous spectrum or ‘continuum’, a bit like someone who listens to different views, then spouts endless platitudes to please everyone.

Where is Population III?
The minute we own something it becomes more valuable to us – we make connections with objects almost instantly and are loathe to give them up. Split a test group into two and give the groups differing objects (it doesn’t matter what: money or a lottery ticket, chocolate or a mug, a mug or a different mug); they will treasure what they have been given, showing a reluctance to trade independent of what they have been given. This is called the endowment effect. There’s something in us that needs to classify objects, own them and keep them. British physicist Ernest Rutherford famously said that ‘All science is either physics or stamp collecting’, implying derision on, for example, the practice in biology of classifying the animal and plant kingdoms into species, genera, families and so on.

The 1890s saw the industrialisation of stellar classification, not with the use of computers as we know them, but with the use of women, who were referred to as ‘computers’. The cheap labour women provided allowed Williamina P. Fleming to get a job as an assistant to the Harvard astrophysicist Edward C. Pickering. Fleming evaluated stellar spectra and assigned them a letter according to the strength of the hydrogen absorption lines.

There is a joke in academia that if a paper has a question in the title, you can save yourself reading all the waffle as the answer will always be ‘no’. In 1981 American astrophysicist Howard Bond could have published a paper that said, ‘Have we found Population III?’, to which the answer would have been ‘no’. Instead he entitled his paper ‘Where is Population III?’, which doesn’t demand either a yes or a no, just a frustrated bewilderment. The search for metal-free stars comprised Bond’s PhD thesis. Over years he had searched more than half the sky for any stars with less than a 1,000th of solar metallicity, the definition of a Population III star at the time. His survey achieved a lot, and found some pretty interesting stars down to 1/500th of the solar metallicity and lower ... but not many. And none beyond that Population III boundary.


The Small Bang

Penzias and Wilson had tried everything obvious, so it’s little wonder that they then began to consider the improbable. Two pigeons had made their home right in the depths of the horn of the antenna, beside the heated observation box. They had coated the inside of the antenna in what Penzias politely referred to as a ‘white dielectric material’. Pigeon excrement. A dielectric material is an insulating material, a poor conductor that could create interference in the radio signal. These birds had proved themselves a tenacious pair as every time Penzias and Wilson searched a new area of the sky, the whole antenna rotated, topping the pigeons about as if on a funfair ride. It can’t have been a comfortable place to roost, but pigeons do reportedly mate for life and the telescope was their home.

The birds were trapped in a box and, according to Wilson, shipped to a pigeon-fancier. The pigeon-fancier looked at them, determined that they were junk pigeons, and let them go. No sooner had the white dielectric material been cleaned from the antenna, than the pigeons returned to their spring-cleaned love nest two days later. That homing instinct really is something, isn’t it?


Bird poop

Don’t feel bad if the idea of a Big Bang is ... ridiculous or even meaningless to you. It was to Einstein, so you can let yourself off the hook.
This cosmic microwave background radiation is everywhere because it pervaded every part of the tiny, hot Universe when it exploded from a singularity. The Universe may have got a lot bigger but the radiation, the afterglow, still pervades it. The photons we measure today still retain that blackbody spectrum, the exact distribution we would expect if the Universe began in a Big Bang. The CMB is there, just as it was in 1964, and continues to be picked up as interference by radio antennas. If you remember the days of analogue television, a decent amount of the static on untuned channels was radiation from the Big Bang.

A Lucky Cloud of Gas
To form a star, the conditions in a cloud of gas must be just right. It is a struggle so closely matched in strength that the event can go on for billions of years. Declaring a draw is not an option in this case, though. One team, gravity, has infinite stamina, and while the opposing pressure produced by internal nuclear reactions provides relief, the nuclear fuel will eventually run out. The star will falter while gravity pushes on relentlessly. We might define a star in the driest of terms: a celestial body where the pressures created by internal nuclear reactions are enough to hold up its own collapse due to gravity. Really, though, a star is just a very lucky cloud of gas.
In theory, the larger the star, the more volatile the radiation pressure at the centre of a star is. The energies are so large that they are unstable and even a small energy change, for example from the accretion of a neighbouring gas cloud, can cause large changes to the equilibrium between the gravitational presssure and radiation pressure. Effectively, the larger the star the more likely it is to exhibit diva-like tendencies under pressure and throw tantrums, suddenly expanding and dissipating or suddenly collapsing. Because this behaviour is unpredictable, there is no fast rule for the maximum mass of a star, but when we study the stars around us, those above 50 solar masses are rare. When it comes to the first stars, though, there is reason to believe that this upper mass limit could have been a lot more relaxed than for current star formation.

The Dark Ages

The flat curves imply that as you observe further from the centre of the galaxy, the stars still move just as fast. The only way this can happen is if the amount of mass is increasing as you go further out, so that the gravitational pull on the stars is maintained. This doesn’t align with our optical view of a galaxy: galaxies have bright centres and diffuse outer halos, not the other way around. The conclusion can only be that there is a colossal amount of matter in the galaxy that we can’t see: dark matter.

Dark matter is everywhere. We are sitting in a great big ball of it right now. We don’t know about it because it doesn’t seem to interact with anything apart from the gravitational force, and its density is so low that on human scales we can’t feel the pull at all. Over our lifetimes, only about 1mg of dark matter will pass through our bodies. This low density is why we have a Keplerian descent in the Solar System and not a flat rotation curve. The dark matter is there, but its density is so low that it makes little difference. While Neptune feels the gravitational force exerted by all the inner planets and all the enclosed dark matter, the latter only makes up the mass of a large rock. On a Galactic scale, though, all that matter adds up. In the outer regions, as you increase the radius you increase the volume massively, encapsulating only a few more visible stars but a huge amount more of dark matter. At that scale we finally see it at work, pulling on those outer stars and making them go faster than we thought possible.

If you study the literature, there are a few concepts that could explain the EDGES signal, the most sensational of which is interacting dark matter, because of the unexpected and far-reaching consequences for another scientific field desperate for data. So while we should allow ourselves to feel the excitement of the possibility, the following should be read with the healthy skepticism of a scientist. It is one of a few ideas, and we need more data before we can present it as a robust theory. We have known about the connection between dark matter and gas and galaxies and stars for a long time, and we have included it in our simulations, so it’s not as if we ignore dark matter completely. Once the gas has coalesced into a star within the web of dark matter, however, we have little need to consider dark matter. It’s on the wrong scale to matter, just as dark matter has little effect within the scale of our Solar System. Or so we thought. What the EDGES result implies is that something is interacting with the gas to make it colder, and the only thing hanging around in the Universe colder than the gas is dark matter.
Try hugging someone sprinting in the opposite direction and you’ll understand why sometimes slower is better when it comes to interactions. The Dark Ages provide the calmest of environments, when the gas particles were moving at their slowest speeds. Therefore if dark matter collisions are most likely in cold environments, this is when they would most likely happen. This neatly explains why dark matter is so unsocial now, yet was so much more interactive in the past.
One of the first papers to propose the excess background theory is infused with the same excitement as the first dark matter theory paper. There is even an exclamation point in the text, which I’m not sure I’ve ever seen in a scientific paper before. It’s simply not done. I love seeing this enthusiasm and humanity in scientific papers, but there are others who would argue against me. They would say that science should be dehumanised, clinical and purely logical. I disagree. When you read the older scientific papers from the 1700s to the twentieth century, this excitement is present, alongside admissions of mistakes or confessions to not knowing where to go next. In the twentieth century, it’s hard to pin down when we lost this. This lack of admission of humanity has led to a constant inflation of the success of results, a grandiosity that leads to the smallest of results lauded as ground-breaking. It detracts from collaborative thought and reduces science to branding. I don’t like it.

Fragmenting Stars
The mass that a system must reach to collapse is defined through a quantity called the Jeans mass, named after British astronomer Sir James Hopwood Jeans. [...] Jeans realised that the mass at which a system would become gravitationally bound would depend on how big and how hot the system was. A colder cloud will collapse more readily than a hot cloud because the atoms are less energetic and can less easily zip away. A denser cloud will more readily collapse than a more diffuse cloud because it squeezes the atoms into a closer space.

Stellar Archaeology
I read all about Einstein’s special relativity: how time slows down the faster you move and how a ruler appears shorter if it moves at speeds approaching the speed of light. I was astonished that I hadn’t heard of this before – and didn’t understand a word of it. In only one sitting, it was enough to pull me away from my pursuit of the past into something that sounded like science fiction. I went to university to study physics just so I could understand that book.

It’s easy to look back in hindsight and cherry-pick these kinds of statements from the unassuming papers that changed our world view, but what we are looking at are good scientists. They had a sample size of two metal-poor stars against the accepted knowledge that stars were all the same, all like our Sun, and while a scientist should be confident in their analysis to publish, rarely does a eureka moment come so clearly. More often, instead of ‘Eureka! I’ve got it!’, it is ‘That looks weird! Hang on, what is that? I think that might be ... huh. Hey, can you all check this to make sure I haven’t done something stupid?’ It often requires a few eureka moments for people to be comfortable leaving their comfy, but wrong, theories. [...] Exploring the different velocity populations further, what Roman noticed was that when she divided the sample into weak and normal metal lines, it was only in the weak group that she saw the high velocities. Roman had discovered that stars with more metals tended to be found dawdling about in the disk of the Milky Way. In contrast, more metal-poor stars tended to be found in high-velocity elliptical paths that extrapolated to them residing often all the way out in the halo.




Galactic Cannibalism
Firstly, dark matter is a notoriously tricky entity to track down or constrain. We know it is there from kinematic arguments provided by galaxy rotation curves and galactic mass measurements. Finding it in terms of its particle interactions, and therefore defining what it is, has eluded us, however. To have found what appears to be the highest concentration of dark matter that we have ever detected, and right next door to us, has sent particle physicists all abuzz. The second reason that Segue 1’s definition as a galaxy is very exciting is that it is small. Very small. And that is interesting because the path to becoming a large galaxy is through galactic cannibalism. Galaxies consume each other, so that over cosmic time we go from a Universe full of lots of little galaxies to one with much larger galaxies. If we’ve found a small galaxy, then we may have found a survivor, an ancient morsel cast aside by a hungry Milky Way. We’ve found a first galaxy, and within an aged first galaxy might just lie a lingering first star.
It is unlikely, then, that there was a first galaxy consisting entirely of first stars. Instead, the first stars lived in tiny halos, before coming together to form the first dwarf galaxies. These galaxies would swiftly become enriched with metals and a host of Population II stars. They would continue star formation and accreting other dwarfs, becoming larger and larger, to form the gigantic galaxies we see around us today. However, not all dwarfs were swept up in the fray. Roughly 5–15 per cent of the first tiny galaxies in the Galactic neighbourhood are estimated to survive intact as fossil galaxies around the Milky Way.

Yes, physicists did call their two main dark matter contenders MACHOs and WIMPs, and found vaguely justifiable acronyms to do so. The effort that must have gone into that.

There were particle physicists who wanted to probe the dwarf galaxies for dark matter. There were stellar achaeologists, excited to have found many more focused locations in which to look for their prized metal-free star. Galaxy formation astronomers were there to use these tiny building blocks to understand how the huge galaxies we see today are formed. And there were other early Universe astronomers like myself, who suggested that younger star-forming dwarf galaxies could be used as analogues of the earliest star-forming galaxies in our Universe. [...] The excitement of finding the first fossil galaxy had waned, and the astronomers were already looking to other candidates and planning more observations with glee.

The Cosmic Dusk

The mass of a star determines how quickly it fuses its fuel, and therefore how long it survives. The mass is also important for determining how a star will die.

Below 8 solar masses: white dwarf For the tiny stars, as the stellar core finishes hydrogen and helium burning, it finds itself at too low a temperature to ignite carbon burning. The core undergoes a collapse and the outer layers are expelled gently into what we call a planetary nebula. The core does not collapse completely and is held up by the pressure of all the electrons in the core whizzing about. [...]

Between 8–20 solar masses: neutron star For a more massive star, due to the accompanying increase in internal pressure the stellar core can reach higher temperatures, igniting all nuclear burning stages, for example of carbon, oxygen and silicon. Once silicon burning turns the core to iron, the core collapses, as it cannot fuse iron to create energy – the process uses more energy than it creates. With a greater mass than a white dwarf, the core can continue collapsing even in the face of the pressure generated by the electrons. In the core, the sea of protons and electrons is squeezed to form neutrons, and the pressure they exert is what eventually stops the collapse. The sudden bounce back when the neutrons are formed causes a shock wave to radiate outwards, expelling the outer layers of the star in an explosion called a supernova. This leaves a neutron star, a stellar remnant that will cool and fade. [...]

Between 20–100 solar masses: black hole Black holes are formed when the neutron pressure within the core can no longer hold up the outer layers of a star. The collapse continues, compressing the matter so much that it produces an extremely strong gravitational field. Within a certain proximity, called the event horizon, nothing can escape – not dust, not metals, not gas and not even light. [...]

Black holes are the bits of our Universe where the standard physics we know and love has given up and gone for a nap, leaving us on the sofa with a kooky, unpredictable aunt in charge, the kind that ‘has no rules’.


Black holes

Between 100–260 solar masses: pair-instability supernova In this mass range, the core temperature of a star reaches such high levels after helium burning that it converts some photons into electron-positron pairs in a process called pair production. [...] When pair production happens, some photon pressure pushing outwards is lost. The sudden loss of pressure results in a rapid core collapse, as the outer layers rush downwards at the mercy of gravity. For stars below 140 solar masses, there results a pulsing behaviour, as a series of mass shells explodes outwards. For stars above 140 solar masses (and below 260 solar masses), the downwards crush of material is enough that rapid nuclear burning of oxygen and silicon during the core collapse produces an explosion that disrupts the entire being of the star. The whole thing explodes, leaving no remnant whatsoever, no gravestone or even tiny wooden cross. [...]

Above 260 solar masses: direct collapse black holes (DCBHs) These, currently theoretical, entities result from a gas cloud being in the right place at the right time. While most clouds are cooling via molecular hydrogen and forming protostars, there will remain some clouds that haven’t quite got to that stage. If such a cloud is situated next to another star-forming site, the radiation from the new stars can flood the cloud, breaking up molecular hydrogen. Within these neighbouring dark halos, molecular hydrogen is destroyed, leaving the cloud unable to cool down, keeping the Jeans mass high and suppressing fragmentation. The whole cloud then collapses as one into an enormous protostar. [...]

The first stars are likely to have ended their lives as pair-instability supernovae, or they might even have collapsed directly into a black hole early in the protostar formation. We have two upcoming methods of peering into the Cosmic Dawn, or rather the Cosmic Dusk as the first stars set for the last time. The first, a folding infrared telescope hoping to see the supernovae, the second, a space triangle that might shiver as the gravitational waves from colliding DCBH pass by. Both require precision planning and testing, and the funding necessary for that. They are both the types of experiment that will leave astrophysicists sick with nerves as they make their lengthy journeys and unfurl in space. But with that sense of risk come huge rewards. Without the atmosphere of the Earth in the way, they can peer further and into smaller corners of the Universe, uncovering the first stellar deaths, if not the first stars themselves. It’ll be like all our Christmases have come at once.


The Epoch of Reionisation
The first stars. Hidden black holes. Looking back in time. All of these possibilities sound wonderful. So the fact that they all come under the umbrella of a subject called the ‘Epoch of Reionisation’ is, well, an anti-climax. What a terrible, terrible name. It took me several months to spell it confidently, let alone pronounce it, and trust me when I say that no one turns up to your public lectures if you advertise it as the topic. Change the title to ‘The dark ages of the Universe and the first stars’, however, and you’re in the money.*

* Figuratively. Early on in my career the most I had been paid for a public lecture was a doughnut, and grateful I was for that too.

As the Universe expanded and cooled, the electrons fell into company with the drifting protons, forming hydrogen atoms. The photons could travel unimpeded. We say that the Universe became transparent to photons, and formed what we can measure today as the cosmic microwave background. I imagine a train concourse or plaza full of adults (protons) and their hyperactive children (electrons) running about in all directions. As a bystander, it is almost impossible for you (a photon) to walk through to the other side in a straight line, without dodging or diverting. When the adults take control and grab each child by the hand, however, it becomes much easier to walk through without deviating. [...] The early Universe was filled with neutral hydrogen. Over only a few hundred million years this hydrogen was either wrapped up into the first stars or ionised, returned to its constitutent state of separate electrons and protons. We call this change in the state of the Universe’s hydrogen, from neutral to ionised, the Epoch of Reionisation, returning as it did to the form it was in right in the beginning at the Big Bang.

Compare to:

Marcelo Gleiser

That we make conclusions or judgements based on evidence is so intuitive that any other possibility seems ridiculous in the present day. There are a lot of distinct possibilities for how these ionised bubbles will look and grow, and we need to narrow down the possibilities based on the evidence. It’s a classic whodunnit, or should I say whatdunnit.



Astronomy is time travel, and while optical astronomy is the Back to the Future DeLorean, radio astronomy is Dr Who’s TARDIS, taking us to times and distant places in our Universe unreachable by any other means.


Time Travel

Originally, the military had wanted to develop a ‘Black Box’, a device that would shoot a concentrated beam of radio waves at an enemy plane and simply blast it out of the sky. This was proved unfeasible,* and attention was turned to developing a radar system small enough to sit in a fighter plane.

* A £1,000 reward from the Air Ministry for anyone who could kill a sheep at a distance of 100 yards went unclaimed.

A petabyte is equivalent to about 2000 years worth of MP3 songs.

Some scientific fields (no, I won’t say which) are famous for being exceedingly competitive, with sniping atmospheres and confrontational conferences, but mine ... is just lovely. Despite most of us being in some telescope team or another, the competition is the healthiest it can be – not that criticisms and questions are not aired, but they are often followed by an offer of help or a workable solution. We’re a small field but we all get on rather well, which is a good thing really considering we will probably find ourselves in the same room at least twice a year for the rest of our working lives.

Unknown Unknowns
Sometimes having a job that reminds you daily how insignificant you are can be ... wearing. Largely, however, my connection with cosmology has been positive. It makes me feel lucky. I know what a balancing act the path to our existence has been and the serendipity that has helped us to understand the Universe to the level that we do. We’ve been so resourceful that a piece of rock orbiting our planet has become the doorway to understanding our ancient past, a past we’re on the verge of seeing for ourselves. Somehow, these tiny, insignificant organisms have been able to stop fighting among themselves for long enough to plan for an experiment decades away. I feel as though I’ve been given tickets to the world’s greatest spectacle. That’s not depressing, that’s amazing. Time to enjoy the show.

text checked (see note) May 2021

top of page

Basic "stars" background graphic copyright © 2003 by Hal Keen

The strip of celestial objects along the left side are strung-together excerpts from the Hubble Ultra Deep Field image.