OUR COSMIC HABITAT II: BEYOND OUR GALAXY
Telescope (n): A device having a relation to the eye similar to that of a telephone to the ear, enabling distant objects to plague us with a multitude of needless details.
Ambrose Bierce
THE UNIVERSE OF GALAXIES
I’ve described how the atoms of the periodic table are made: that we’re stardust – or, less romantically, the ‘nuclear waste’ – from the fuel that makes stars shine. These processes depend on the strength of the ‘nuclear force’ that glues together the protons and neutrons within the nuclei of these atoms – measured by the cosmic number ε = 0.007 that denotes the proportion of energy that is released when hydrogen fuses into helium. But where did the original protons and hydrogen atoms come from, and how did the primordial material aggregate into the first galaxies and stars? To answer these questions, we must extend our horizons in space and time – out to the extragalactic realm, and back to an era before the birth of the first stars. We shall encounter further numbers that describe our entire universe, and discover that our emergence depended on these too being finely tuned.
Stars are agglomerated into galaxies, which are the basic units that make up the universe. Our own is typical. Its hundred billion stars lie mainly in a disc, circling around a bright inner ‘bulge’ where the stars are closer together than average. Right at the centre lurks a black hole with the mass of 2.5 million suns. A light signal would take about 25,000 years to reach us from the galactic centre, and we on Earth are rather more than halfway out towards the disc’s edge. From our Sun’s location, the other stars in the disc appear concentrated in a band across the sky, known to us as the Milky Way. Typical stars take more than a hundred million years for a single circuit (sometimes called a ‘galactic year’) around the galactic centre.
Andromeda, our galaxy’s nearest major neighbour in space, is about two million light-years away. To an astronomer on a planet orbiting one of Andromeda’s stars, our galaxy would look rather like Andromeda does to us: a disc, viewed obliquely, made of stars and gas circling around a central ‘hub’. Millions of other galaxies are visible with large telescopes. Not all are disc-like: the other important class is the so-called ‘elliptical galaxies’, in which the stars are not organized into a disc but are swarming around in more random orbits, each feeling the gravitational pull of all the others.
Galaxies are not sprinkled around randomly in space: most are in groups or clusters, held together by gravity. Our own Local Group, a few million light-years across, contains the Milky Way and Andromeda, together with thirty-four smaller galaxies (that, at least, was the last count – very faint and small members of the Local Group are still being discovered). Gravity is pulling Andromeda towards us at about 100 kilometres per second. In about five billion years, these two disc galaxies may crash together. Such crashes are ‘routine’ cosmic events: we see, deeper in space, many other galaxies that seem to be undergoing such an encounter with another.
Galaxies are so vast and diffuse, and the stars are so thinly spread, that actual collisions between individual stars are exceedingly rare. (This is clearly true in the Solar neighbourhood, because even the nearest stars seem like faint points of light). Even when two galaxies crash together and merge, there would be very few stellar impacts. All that happens is that each star feels the collective gravity of everything in the other galaxy. Orbits are so distorted that the stars end up in a single chaotic swarm rather than two separate discs. This is, of course, just what a so-called elliptical galaxy looks like, and I suspect (though the issue is still controversial) that the big elliptical galaxies were formed in this fashion.
THE TEXTURE OF OUR UNIVERSE: THE COSMIC WEB
Our Local Group is near the edge of the Virgo Cluster, an archipelago of several hundred galaxies, whose core lies about fifty million light-years away. The clusters and groups are themselves organized into still larger aggregates. The so-called ‘Great Wall’, a sheet-like array of galaxies about 200 million light-years away, is the nearest and most prominent of these giant features. Another concentration of mass, the ‘Great Attractor’, exerts a gravitational force that pulls us, and the entire Virgo Cluster as well, at several hundred kilometres per second.
Many phenomena in nature – mountain landscapes, coastlines, trees, blood vessels, and so forth – are ‘fractals’. A fractal is a pattern with the special mathematical feature that a small part, when magnified, resembles the whole. If our universe were like this – if it contained clusters of clusters of clusters . . . ad infinitum – then however deeply we probed into space, and however large a volume we sampled, the galaxies would still have a patchy distribution: by probing deeper, we’d simply be sampling larger and larger scales in the clustering hierarchy. But this is not how our universe looks. Powerful telescopes reveal galaxies out to several billion light-years. Within this far larger volume, astronomers have mapped many more clusters like Virgo, and more features like the ‘Great Wall’. But deeper surveys don’t reveal any conspicuous features on still larger scales; in the words of the Harvard astronomer Robert Kirshner, we reach ‘the end of greatness’. A box whose sides are 200 million light-years (a distance still small compared with the horizon of our observations, which is about 10 billion light-years away) is capacious enough to accommodate the largest structures, and to contain a ‘fair sample’ of our universe. Wherever it is placed, such a box would contain roughly the same number of galaxies, grouped in a statistically similar way into clusters, filamentary structures, etc. The hierarchy of clustering doesn’t continue towards indefinitely large scales.
Our universe is thus not a simple fractal; moreover the ‘smoothing scale’ is small compared with the largest distances that our telescopes can probe. As an analogy, imagine you were on a ship in the middle of the ocean. A complicated pattern of waves would surround you, stretching to the horizon. But you could study the statistics of the waves because your field of view extends far enough to encompass many of them. Even the biggest waves on the ocean are far smaller than the horizon distance, and you could, in your imagination, divide what you can see into many separate patches, each large enough to be a fair sample. There is a contrast here between seascapes and landscapes: in mountainous terrain, one grand peak often dominates the entire horizon and you can’t define meaningful averages as you can for a seascape. (Landscapes, indeed, can be fractal-like. The mathematics of fractals is used in computer graphics programs for depicting imaginary landscapes in movies.)
Cosmic structures encompass a wide range of dimensions: stars, galaxies, clusters, and superclusters. On scales less than about 1/300 of the horizon, the concentration of galaxies varies by more than a factor of two from place to place; on larger scales, the fluctuations are gentler (though there are a few conspicuous features like the Great Attractor). Superclusters of galaxies – to extend the ocean analogy – are like the longest conspicuous waves. We shall see in Chapter 8 that this scale depends on a single cosmic number, Q imprinted in the very early universe, and that the ‘embryos’ of clusters and superclusters – structures stretching millions of light-years across the sky – can be traced back to a time when the entire universe was of microscopic size. This is perhaps the most astonishing link between the outer space of the cosmos and the inner space of the microworld.
One’s first guess might be that the texture of our universe on such large scales was irrelevant to our local habitat within the Solar System: it might not seem to matter whether our galaxy contained a quadrillion stars, or else ‘only’ a million, rather than the hundred billion that we observe; nor whether it belonged to a cluster containing millions of other galaxies rather than just a few. But a universe much rougher than ours wouldn’t be hospitable to stars and planets. On the other hand, a universe that was too smooth would be blandly uninteresting: no galaxies and stars would form, and all the material would be thinly spread and amorphous.
This will be the theme of Chapter 8. But, for the moment, we can note another crucial consequence of the large-scale smoothness: it makes the science of cosmology possible, by allowing us to define the average properties of our universe – the demography of the galaxies, the statistics of the clusters, and so forth. Despite galaxies and clusters, it is still useful to think about the smoothed-out properties of the universe, just as we can describe the Earth as ‘round’, despite the complex topography of its mountains and its ocean depths. However, it would not be useful to describe the Earth as ‘basically round’ if its mountains were thousands of kilometres high rather than just a few.
Even more important, we can meaningfully ask whether our entire universe is static, expanding or contracting.
THE EXPANSION
Galaxies are the ‘building blocks’ of our universe, and by studying the light from them we can infer how they are moving. The hundred billion stars in a typical galaxy are too faint to be seen individually: telescopes record the total light from many stars blurred together. This light can be analysed into a spectrum. We have noted how the light from a single star can reveal its speed towards (or away from) us, and how repeated measurements can even pick up the tiny oscillatory motion induced by an orbiting planet. Likewise, the spectrum of an entire galaxy reveals how fast it is moving, either towards us (a shift towards the blue end of the spectrum) or away from us (a shift towards the red).
Perhaps the most important single fact about our universe is that the light from all distant galaxies is shifted towards the red: all (except for a few nearby galaxies in the same cluster as our own) are receding from us. Moreover, the redshift (a measure of the recession speed) is larger for the more distant galaxies. We seem to be in an expanding universe, where clusters of galaxies get more widely separated – more thinly spread through space – as time goes on.
The simple relation between redshift and distance is named after Edwin Hubble, who first claimed such a law in 1929. Observers on other galaxies would witness a similar expansion of distant regions away from them. The expansion is a broad-brush effect: individual galaxies (even clusters of galaxies) are not themselves expanding; still less does the expansion affect anything more local, such as our Solar System.
Imagine that the rods in the M. C. Escher drawing in Figure 5.1 lengthened at the same rate. An observer on any vertex would see the others receding at speeds that depended on how many intervening rods there were. In other words, the recession speed of other vertices would be proportional to their distance. The galaxies aren’t in a regular lattice – as already mentioned, they are in groups or clusters – but you can nonetheless envisage the expansion by imagining that clusters of galaxies are linked by rods that all lengthen at the same rate. There is nothing special about any vertex in the picture; and there is likewise nothing special about the location of our galaxy in the universe. (Although our galaxy is randomly placed, we are not, however, observing it at a random time; the reasons for this will become clear later.) Cosmology has only progressed because our universe, in its large-scale structure, is uniform enough to be described by a simple ‘Hubble expansion’, where all patches seem to be expanding similarly. The expansion can be envisaged locally as a Doppler effect, but on large scales, when the apparent recession is at a good fraction of the speed of light, it is better to attribute the redshift to a ‘stretching’ of space while the light travels through it. The amount of reddening – in other words, the amount that the wavelengths are stretched – is then equal to the amount by which the universe has expanded (and, in our Escher analogy, the ‘rods’ have lengthened) while the light has been travelling towards us.
We might, of course, wonder whether the redshift actually implies expansion, rather than some new physical effect that comes into play over long distances. The possibility of such a ‘tired light’ effect is still sometimes raised, although nobody has come up with a viable theory consistent with all the evidence (it must, for instance, produce the same fractional change in wavelength for light of all colours, and mustn’t blur the images of distant objects). A non-expanding universe would actually entail even worse paradoxes than any Big Bang theory. Stars don’t have infinite energy reserves; they evolve, and eventually exhaust their fuel. So therefore do galaxies, which are essentially aggregates of stars. It is possible to date the oldest stars in our Milky Way, and in other galaxies, by comparing their properties with the outcome of computations of how stars evolve. The oldest are about ten billion years old – entirely consistent with the view that our universe has only been expanding for a bit longer than that. If our universe were static, all galaxies must have mysteriously ‘switched on’ in their present positions – in a synchronized fashion – about ten billion years ago. A non-expanding universe would entail severe conceptual difficulties.
The expansion almost certainly began between ten and fifteen billion years ago, twelve or thirteen billion being the best guess. There are two reasons for this persistent uncertainty about the age of our universe. The exact distances to galaxies are (unlike their recession speeds) still somewhat inexact; also, the estimate depends on how much faster (or slower) the expansion might have been in the past.
SEEING INTO THE PAST
Light travels at a finite speed, and so we see distant regions not as they are now but as they were a long time ago. At earlier epochs, the universe would have been more compressed – the rods in our lattice would have been shorter. So the second Escher picture, Angels and Devils, shown in Figure 5.2, better represents what we actually see.
We’d expect very distant galaxies to look different from nearby ones. Their light has taken a long time on its journey, and so they were younger and less evolved when they emitted the light now reaching us. Not all the pristine gas had at that stage condensed into stars. These evolutionary changes would be so slow that they would only be manifest over billions of years. To detect a trend, one must therefore probe galaxies so far away that their light set out several billion years ago.
The Hubble Space Telescope (HST) – named in honour of the discoverer of cosmic expansion – circles the Earth far above the blurring effect of the atmosphere and has produced the sharpest pictures yet of very distant regions. The HST is so sensitive that a long exposure reveals, close packed in the sky, literally hundreds of faint smudges, even within a field of view so small that it would cover less than a hundredth of the area of the full moon – and would appear as a blank patch of sky when viewed with an ordinary telescope. (I think that the amazing pictures being generated by the HST will impact as strongly on public consciousness as the first images from space, in the 1960s, that showed the whole Earth, with its delicate-seeming biosphere.) The faint features in these pictures, with a diversity of shapes, are a billion times fainter than any star we can see with the unaided eye. But each is an entire galaxy, thousands of light-years in size, which appears so small and faint because of its huge distance from us. These galaxies look different from their nearby counterparts because they are being viewed when they have only recently formed: they have not yet settled down into steadily spinning discs like the photogenic nearby spiral galaxies depicted in most astronomy books. Some consist mainly of glowing diffuse gas, which hasn’t yet fragmented into stars. Most of them appear intrinsically bluer than present-day galaxies (after correcting, of course, for the redshift), because massive blue stars, which would all by now have died, were still shining when the light left these distant galaxies.
These very deep images show us what a galaxy like our own Milky Way would have looked like when its first stars were shining brightly. When we observe Andromeda, a nearby ‘twin’ of our own galaxy, we may wonder whether Andromedans are looking back at us with still bigger telescopes. Perhaps they are. But there could be nothing so ‘advanced’ on these very remote galaxies: we are viewing them at a very primitive evolutionary stage, before enough time has elapsed for many stars to have completed their lives. They have as yet no complex chemistry; there is very little oxygen, carbon, etc, even to make planets; and so there is scant chance of life. We are seeing these galaxies at a stage when the basic building blocks for planetary systems were being laid down. (The light we detect was actually emitted in the far ultra-violet. Such radiation cannot be detected by the eye, nor can it even penetrate the Earth’s atmosphere. But the extreme ultra-violet radiation from these galaxies, by the time it reaches us, has been shifted into red light.)
The most distant galaxies are so redshifted that the wavelength of the light has been stretched by more than a factor of six: that’s how much the universe must have expanded since the light set out. If the expansion had been steady, with galaxies neither accelerating nor decelerating, then when the universe was one-sixth its present scale (in the sense that distances – the rods on Escher’s lattice – were scaled down six times smaller) it would have been one-sixth its present age. This statement might at first seem troublesome: doesn’t it mean that a galaxy must be moving away at five times the speed of light, if the light has taken five-sixths of our universe’s present age to get back to us? But there’s no paradox. Einstein’s special theory of relativity tells us that nothing can move faster than light, relative to us, when time is measured by our clock. But that theory also tells us that a fast-moving clock runs slow. A fast clock can indeed travel five light-years for every year that it records if it moves at ninety-eight per cent of the speed of light.
The situation is actually a bit more complicated because the recession speed would not be constant. The gravitational pull that everything in the universe exerts on everything else causes deceleration, which tends to make the earlier stages of cosmic expansion relatively even shorter. But (as discussed in Chapter 7) another force may be at work that tends to speed up the expansion. There is still, therefore, some uncertainty about how far back in time (or how far away in space) these remote galaxies actually are: the best guess would be that their light set out when the universe was around one-tenth of its present age.
Cosmologists study ‘fossils’ of the past: old stars, chemical elements synthesized when our galaxy was young, etc. In that respect they are like geologists or palaeontologists trying to infer how our Earth and its fauna have evolved. But cosmologists actually have an edge over other scientists who can’t do experiments and depend on ‘historical’ evidence. By directing their telescopes towards distant objects, cosmologists can see the evolution they claim: populations of distant galaxies, whose light set out several billion years ago, look different from their counterparts nearby. Because of the large-scale uniformity, all parts of the universe have had similar histories. These remote galaxies should therefore – statistically at least – look similar to the way our Milky Way, the Andromeda galaxy and other nearby systems would have looked billions of years ago.
The field of view of a telescope is a long thin cone, extending out to the limits of vision. The objects at each distance tell us about a specific epoch in the past. As we probe greater distances, we probe further back in time, just as a borehole through successive layers of Antarctic ice can reveal the history of the Earth’s climate.
The Hubble Space Telescope was dogged by delays, errors and cost overruns, but it has now, albeit belatedly, fulfilled the hopes that astronomers had for it. Its out-of-focus mirror was corrected by the first manned repair mission in 1994; and the light detectors on board have been upgraded. It could, barring mishap, continue until 2010, by which time still larger space telescopes may have been deployed. But equally important has been the advent of a new generation of larger telescopes on the ground. Their 8–10 metre mirrors offer sixteen times more collecting area than the HST, and so they can collect far more light from very faint and remote galaxies. The two Keck Telescopes, on Mauna Kea in Hawaii, were the first of these new-generation instruments to come into service, but there are now several more. Most impressive of all is the Very Large Telescope (VLT), a connected cluster of four telescopes, each with an eight-metre mirror, constructed in the Chilean Andes by a consortium of European nations.
The sharpness of the images from these ground-based telescopes is limited by the blurring due to turbulence in the atmosphere (the same process that makes stars twinkle). This limit can be surmounted either by linking two or more telescopes together and combining the images, or by so-called ‘adaptive optics’, whereby a mirror is continually tweaked and adjusted to compensate for fluctuations in the atmosphere.
These superb instruments offer snapshots of the universe right back to when the first galaxies were forming. The first stars may actually have formed even earlier, in aggregates smaller than present-day galaxies, but which are too faint for us to see. These later agglomerated into larger structures. The rate at which gas condenses into stars is the ‘metabolic rate’ of a galaxy. It seems to have peaked when the universe was about a quarter of its present age (even though the very first starlight appeared much earlier). Fewer bright stars are forming now because most of the gas in ‘mature’ galaxies has already been incorporated into older stars.
That, at least, is the scenario that most cosmologists accept. Fleshing out the details will need more observations and a fuller understanding of how stars form. The aim is to obtain a consistent scenario that not only matches all we know about present-day galaxies but also takes into account the increasingly detailed snapshots of what they looked like, and how they were clustered, at all earlier times. When data are sparse, they may all fit with several completely wrong theories; but as the evidence mounts up, we should ‘home in’ on a single picture of how things work.
With increasing distance our knowledge fades and fades rapidly. Eventually we reach the dim boundary, the utmost limits of our telescope. There we measure shadows, and we search among ghostly errors of measurement for landmarks that are scarcely more substantial. The search will continue. Not until the empirical resources are exhausted need we pass on to the dreamy realm of speculation.
These are the concluding words of Edwin Hubble’s classic (1936) book, The Realm of the Nebulae. Recent progress would have delighted, and probably astonished, Hubble. That progress is owed to the telescope in space that bears his name, and huge new telescopes on the ground.
BEFORE THE GALAXIES
What about still earlier epochs, before any galaxies could have formed? The best evidence that everything really emerged from a dense ‘beginning’ is that intergalactic space isn’t completely cold. This warmth is an ‘afterglow of creation’. It manifests itself as microwaves, the kind of radiation that generates heat in a microwave oven but very much less intense. The first detection of the ‘cosmic microwave background’, back in 1965, was the most important advance in cosmology since the discovery of the expansion of the universe. Later measurements confirmed that these microwaves have a very distinctive property: their intensity at different wavelengths, when plotted on a graph, traces out what physicists call a ‘black body’ or ‘thermal’ curve. This particular curve is expected when the radiation has been brought into balance with its environment (as happens deep inside a star, or in a furnace that has burnt steadily for a long time); it’s just what would be expected if the microwaves were indeed a relic of a ‘fireball’ phase when everything in our universe was squeezed hot, dense and opaque.
By far the most precise measurements came during the 1990s from NASA’s Cosmic Background Explorer Satellite (COBE). When experimenters present their results, they conventionally draw ‘error bars’ indicating the range of uncertainty, but for the COBE data the ‘bars’ can’t be exhibited because they would be shorter than the thickness of the curve. This truly remarkable measurement, with an accuracy of one part in 10,000, confirms beyond reasonable doubt that everything in our universe – all the stuff that galaxies are now made of – was once a compressed gas, hotter than the Sun’s core.
The present average temperature of the universe is 2.728 degrees above absolute zero. This is, of course, exceedingly cool (around – 270°C); but there’s a well-defined sense in which intergalactic space still contains a lot of heat. Every cubic metre contains 412 million quanta of radiation, or photons: in comparison, the average density of atoms in the universe is only about 0.2 per cubic metre. This latter number is less precisely known, because we are unsure how many atoms may be in diffuse gas or ‘dark’ matter, but there seem to be about two billion photons for every atom in the universe. During the expansion of the universe, the density of atoms and of photons both decrease. But the decrease is the same for both, and so the ratio of photons to atoms stays the same. Because this ratio of ‘heat’ to ‘matter’ is so large, the early universe is often referred to as a ‘hot’ Big Bang.
The hot early phases wouldn’t have lasted long. Only for a few minutes would the temperature have exceeded a billion degrees. After about half a million years it had cooled to 3000 degrees – a bit cooler than the Sun’s surface. This marks a significant stage in the expansion process: before that time, everything was so hot that electrons were dislodged from nuclei and moved freely; but afterwards the electrons would have slowed down enough to attach themselves to nuclei, forming neutral atoms. These atoms cannot scatter the radiation as efficiently as free electrons were able to during the earlier and hotter stages. The primordial material would thereafter have been transparent; the ‘fog’ would have lifted. During expansion, the temperature drops inversely with the scale of the universe (the length of the rods on Escher’s lattice). The microwaves that COBE detects are relics from the era when our universe was more than a thousand times more compressed – at 3000 degrees rather than 2.7 degrees, and long before any galaxies came into existence. The intense radiation in the original fireball, although cooled and diluted by expansion, still pervades the whole universe.
The often-used analogy with an explosion is misleading inasmuch as it conveys the image that the Big Bang was triggered at some particular centre. But as far as we can tell, any observer – whether on Earth, on Andromeda, or even on the galaxies remotest from us – would see the same pattern of expansion. The universe may once have been squeezed to a single point, but everyone had an equal claim to have started from that point; we can’t identify the origin of the expansion with any particular location in our present universe.
It is also incorrect to think of the high pressure in the early universe ‘driving’ the expansion. Explosions are caused by an unbalanced pressure; bombs on Earth, or supernovae in the cosmos, explode because a sudden boost in internal pressure flings debris into a low-pressure environment. But in the early universe the pressure was the same everywhere: there was no edge, no ‘empty’ region outside. The primordial gas cools and dilutes, just as happens to the contents of an expanding box. The extra gravity due to the pressure and heat energy actually slows down the expansion.1
This is a consistent picture, but it leaves some mysteries. Above all (since the explosion analogy is flawed) it offers no explanation for why expansion occurs at all. The standard Big Bang theory simply postulates that everything was set up with just enough energy to go on expanding. An answer to why it is expanding at all must be sought in the still earlier stages, where we don’t have such direct evidence nor such a confident understanding of the physics.
The name ‘Big Bang’ was introduced in the 1950s by the celebrated Cambridge theorist Fred Hoyle (already mentioned in Chapter 4 for his insights into the origin of carbon) as a derisive description of a theory he didn’t like. Hoyle himself favoured a ‘steady state’ universe, in which new atoms and new galaxies were imagined to form continuously in the gaps as the universe expanded, so that its average properties never changed. There was at that time no evidence either way – cosmology was the province of armchair speculators – because observations didn’t probe far enough for the evolution (if it existed) to show up. But the steady-state theory fell from favour as soon as evidence emerged that the universe was actually different in the past. Though it turned out wrong, the steady-state theory was a ‘good’ theory in that it made very clear-cut and testable predictions; it was a genuine stimulus to the subject, goading observers to push their techniques to the limit. (A ‘bad’ theory, in this sense, is one that is so flexible that it can be adjusted to account for any data. The eminent – and arrogant – physicist Wolfgang Pauli would deride such vague ideas as ‘not even wrong’.) Hoyle himself never became fully reconciled to the Big Bang, although he adopted a compromise picture that sceptical colleagues called a ‘Steady Bang’.
NUCLEAR REACTIONS IN THE BIG BANG
According to the Big Bang theory, our universe started off hotter than the centre of a star. Why, then, weren’t the primordial nuclei of hydrogen all transmuted into iron during the Big Bang? (Remember that nuclei of iron are more ‘tightly bound’ than any others, and are built up in the cores of the biggest and hottest stars.) If this had happened, no long-lived stars could have existed in our present universe, because all the available fuel would have been used up in the early fireball: a star made of vaporized iron could exist, but it would deflate within millions of years, instead of billions, rather as Kelvin thought the Sun would. Fortunately, the first few minutes of the expansion didn’t allow enough time for nuclear reactions to ‘process’ any of the primordial material into iron – nor even into carbon, oxygen, etc. The reactions would turn about twenty-three per cent of the hydrogen into helium, but (apart from a trace of lithium) no elements higher up the periodic table emerge from the Big Bang itself.
This primordial helium is, however, crucial and offers us strong corroboration of the Big Bang theory. Even the oldest objects (in which pollution by carbon, oxygen and so forth is a hundred times less than in the Sun) turn out to contain 23–24 per cent of helium: no star, galaxy or nebula has been found where helium is less abundant than this. It seems as though the galaxy started not as pure hydrogen, but was already a mix of hydrogen and helium. (The Sun’s outer layers have twenty-seven per cent helium, the extra 3–4 per cent being just about what would have been made, along with carbon, oxygen and iron, in the short-lived early stars that must already have polluted the cloud from which our Solar System formed.)2
Many slow-burning low-mass stars survive, which formed several billion years before our Sun when our galaxy was young. These contain far less carbon, oxygen and iron relative to hydrogen than the Sun does – something that is, of course, natural if, as Hoyle was first to argue, these atoms were expelled from massive stars and accumulated gradually over galactic history. Hoyle’s view contrasted with George Gamow’s idea that the entire periodic table was ‘cooked’ in the early universe. If Gamow had been right and these elements predated the first stars and galaxies, their abundance would be the same everywhere, in young and old stars alike.
Helium is the only element that, according to calculations, would be created prolifically in a Big Bang. This is gratifying because it explains why there is so much helium and why the helium is so uniform in its abundance. Attributing helium to the Big Bang thus solved a long-standing problem, and emboldened cosmologists to take the first few seconds of cosmic history seriously.
As a bonus, the Big Bang accounts for another kind of atom: deuterium (also known as ‘heavy hydrogen’). An atom of deuterium contains not just a proton but a neutron as well, which adds extra mass but no extra charge. The existence of deuterium is otherwise a mystery, because it is destroyed rather than created in stars: as a nuclear fuel it is easier to ignite than ordinary hydrogen, and so newly formed stars would burn up any deuterium during their initial contraction, before settling down in their long hydrogen-burning phases.
Helium and deuterium were made when the temperature in the compressed universe was (in round numbers) three billion degrees – about a billion times higher than it is now. As the universe expands, we can imagine the rods of Escher’s lattice (see Figure 5.1) lengthening. The wavelengths of the radiation stretch in proportion to the length of the rods, and the temperature decreases as the inverse of the length. This means that, when the temperature was around three billion degrees (rather than around 3 degrees as it is now) the rods were a billion times (109) shorter and the densities higher by the cube of that factor, 1027. But our present universe is so diffuse – around 0.2 atoms per cubic metre – that even when compressed by this huge factor, the density is still less than that of air! The temperature was then so high that the individual nuclei would have been in agitated rapid motion. Laboratory experimenters can check what happens when hydrogen and helium nuclei crash together with the same energies as they would have had when the helium-formation occurred, so the calculations are based on quite conventional and firmly based physics.
If we assume a present density of 0.2 atoms per cubic metre, the computed proportions of hydrogen, helium and deuterium that would emerge from the cooling ‘fireball’ universe agree with observations. This is gratifying, because the observed abundances could have been entirely out of line with the predictions of any Big Bang; or they might have been consistent, but only for a density that was far below, or else far above, the range allowed by observation. As we have seen, 0.2 atoms per cubic metre is indeed close to the smoothed-out density of galaxies and gas in our universe. (This has important implications for ‘dark matter’, as discussed in the next chapter.)
