Miércoles 26 de Noviembre de 2008, Ip nº 256

Is Earth at the heart of a giant cosmic void?
Por Marcus Chown

It was the evolutionary theory of its age. A revolutionary hypothesis that undermined the cherished notion that we humans are somehow special, driving a deep wedge between science and religion. The philosopher Giordano Bruno was burned at the stake for espousing it; Galileo Galilei, the most brilliant scientist of his age, was silenced. But Nicolaus Copernicus's idea that Earth was just one of many planets orbiting the sun - and so occupied no exceptional position in the cosmos - has endured and become a foundation stone of our understanding of the universe.

Could it actually be wrong, though? At first glance, that question might seem heretical, or downright silly. But as our cosmic horizons have expanded over the centuries so too has the scope of Copernicus's idea. It has morphed into the Copernican, or cosmological, principle: that nothing distinguishes the position of Earth's galaxy from any other place in the entire universe. And that idea, some cosmologists point out, has not been tested beyond all doubt - yet.

That could be about to change. A new generation of experiments might shore up the cosmic orthodoxy - or blow it out of the water. That unexpected alternative, some people go so far as to say, might be no bad thing at all.

The modern-day Copernican principle amounts to two assumptions. First, that averaged over large enough scales the universe is homogeneous, having essentially the same properties in all locations. Second, that the universe is isotropic, or appears to have the same properties when viewed in any direction from every location. These two ideas are intimately related, but logically separate (see diagram). They were introduced into cosmology not because of any observational evidence, but to save face. In 1917, Albert Einstein had applied his theory of gravity - general relativity - to the dynamics of the universe. Without the simplifying assumptions of homogeneity and isotropy, Einstein's fiendishly complex equations proved impossible to solve.

Even with those assumptions, Einstein's initial insistence that we live in an unchanging universe led him to the wrong solutions. By dropping the "unchanging universe" requirement a few years later, cosmologists created the picture that became the kernel of today's phenomenally successful big bang model. In this picture, the universe started out as a single, infinitely hot and dense point in space, and has since been expanding - initially rapidly, but gradually more slowly as gravity has exerted its pull on the mass of the cosmos.

All seemed well, with evidence in support of the big bang model piling up throughout the 20th century. Then, in 1998, astronomers studying stellar explosions known as type 1a supernovae made a sensational discovery. These supernovae are thought to be uniformly bright, so that the fainter they appear to us, the farther they must be away. But measurements showed that the most distant supernovae did not fit in: they were a lot fainter than they should have been, and seemed impossibly far away. Some time over the past few billion years, they must have begun to race away from us ever faster. Rather than the universe's expansion slowing down, it looked like it was speeding up.

This startling possibility can be accommodated by the standard cosmological equations, but only at a price. That price is introducing dark energy - an unseen energy pervading space that overwhelms gravity and drives an accelerating expansion. Dark energy is problematic. No one really knows what it is. We can make an educated guess, and use quantum theory to estimate how much of it there might be, but then we overshoot by an astounding factor of 10120.

That is grounds enough, says George Ellis, a leading cosmology theorist based at the University of Cape Town in South Africa, to take a hard look at our assumptions about the universe and our place in it. "If we analyse the supernova data by assuming the Copernican principle is correct and get out something unphysical, I think we should start questioning the Copernican principle."

On the face of it, homogeneity and isotropy are unlikely assumptions. Just take a look at the night sky. It is anything but uniform, with most stars concentrated in a band across the sky - the Milky Way.

Of course, that's not the full picture. In 1924, Edwin Hubble discovered that certain diffuse sources of light in the night sky, called spiral nebulae, are actually groups of stars far beyond the Milky Way. The realisation came that the broad swath of the Milky Way is just our own galaxy - the bright lights of downtown seen from our distant suburb - and that it is merely one among countless others splashed across the heavens.

Since then, surveys have shown how galaxies are distributed more or less isotropically - evenly in whichever direction we choose to look. What's more, the cosmic background radiation - the afterglow of the big bang fireball, discovered in 1964 - has pretty much the same intensity and temperature whatever direction we look in.

So while the case for isotropy seems virtually sewn up, the evidence for homogeneity is much less convincing. It is also harder to come by. To create a three-dimensional picture of how matter is distributed in the universe, we need to know how far away different galaxies are. That would mean identifying galaxies that, like type 1a supernovae, are uniformly bright at all distances - a near impossible task, as most galaxies are dynamic, ever-changing bodies.

According to Ellis and others, our uncertainty about galaxy distances allows an interesting possibility. The distribution of matter could look the same in all directions, but vary with distance from us. In particular, we might be sitting in the middle of a "void" - a vast spherical bubble in an otherwise homogeneous universe. This bubble is not devoid of matter. In fact, most of the stars and galaxies we can see from Earth would be contained within it. It's just that everywhere beyond it, which is too far away to see, the density of stars and galaxies is much higher.

How would such a bubble help? In such a low-density region, the braking pull of gravity is weaker, and so the region would quite naturally be expanding faster than the more dense area enveloping it. A bubble surrounding us, covering the volume from which light emitted over the past few billion years is just reaching us, would be just the thing to explain the supernova observations. Observing from within such a bubble, but using distant supernovae as yardsticks, we would see a universe whose expansion seems to be occurring faster than it used to - without the need to invoke dark energy (see diagram).

"Dark energy is a necessity if we assume the supernova acceleration is due to a change in the entire universe's expansion rate over time," says Ellis. "But it's equally possible, and no more radical, to say that it reflects a change in the universe's expansion - in space."

But here's the rub. For things such as the cosmic background radiation to appear isotropic to us from within a void, we would have to be at or extremely close to its centre - which is not only anti-Copernican, but also highly unlikely. Ellis is unperturbed. "We live in an improbable universe," he says. "You can shift around the improbability - for instance, substituting an Earth-centred void for dark energy - but you can't remove the improbability." The problem is to find ways to tell a homogeneous from an inhomogeneous universe. "Without being able to move from our location, that's very hard."

Robert Caldwell of Dartmouth College in New Hampshire agrees. "It would great if there were someone out there who could look back at us and tell us if we're in a void," he says. "Or if we could look in a distant cosmic mirror and see ourselves."

Remarkably, that just might be possible. Caldwell and his colleague Albert Stebbins have been on the case, developing a void-testing idea dreamed up by Jeremy Goodman of Princeton University in 1995 (Physical Review D, vol 52, p 1821). Their scheme involves exploiting the effect that a void would have on the well-travelled photons of the cosmic background radiation (Physical Review Letters, vol 100, p 191302).

The story of these photons starts about 400,000 years after the big bang, when the universe, previously a dense ionised soup of charged nuclei and electrons, had cooled down enough for neutral, uncharged atoms to form. Photons had got stuck in the charged soup, but could now suddenly travel unimpeded through the neutralised cosmos.

Obstacles to the photons' progress began to reappear after some 200 million years, as the first stars or quasars began to re-ionise neutral atoms. Nevertheless, most of these photons continued on untroubled, slowly losing energy as they journeyed through the expanding cosmos. In some cases, they made their way into our telescopes more than 13 billion years down the line.

So what happens to these photons in a void? When they pass by matter, they receive an energy boost. In a void, this gravity assist is less pronounced, and the photons lose energy. They regain the energy on leaving the void again. In fact, because the void itself is expanding and becoming emptier while the photons cross it, they gain a tiny bit more energy on crossing back into a denser region than they lost on entering the void.

As a result of this energy boost, a void would stand out like a sore thumb to most observers, as a colossal "hot" patch in a cosmic background distributed otherwise uniformly across the sky. The only observers not able to see the void in this way would be us earthlings living in the centre of the bubble: all the photons that come our way will have passed through the same amount of void, and so our cosmic background will look completely isotropic. As indeed it does.

But here comes the clever part, say Caldwell and Stebbins. Some of those hotter photons that have passed completely through our void will scatter off ionised gas floating about on the other side and be reflected back our way, as off a mirror. So what we would actually see is a mixture of photons, most of which have come to us directly, but with a smattering of these hotter, reflected photons. As a result, the hotspot in the cosmic background obvious to observers outside our bubble should also be visible to us as a slight distortion in the energy distribution of our cosmic background radiation.

The bigger the void, the hotter the reflected photons and the greater this distortion would be. Current measurements of the cosmic microwave background exclude the effects of voids bigger than about 3 billion light year, as well as voids smaller than about 300 million light years. Caldwell is a sceptic. "We need better data to rule out all possible voids," he says, "but my suspicion is that we don't live in a void and that the Copernican principle will survive."

Ellis, on the other hand, thinks we need more direct ways of distinguishing between a homogeneous universe and an inhomogeneous universe with a void. Three colleagues of his at the University of Cape Town, Chris Clarkson, Bruce Bassett and Teresa Hui-Ching Lu, have something in mind. Their idea exploits gargantuan ripples in the density of matter across the universe, known as baryon acoustic oscillations (Physical Review Letters, vol 101, p 011301). These ripples arose because acoustic waves in the hot matter-and-photon fluid sloshing about in the early universe created regions of higher density, whose greater gravity in turn dragged in yet more matter as the universe expanded.

The researchers propose measuring how the ripples vary in size with distance from us, and therefore at different periods of the universe's evolution. Combined with supernova results, which tell us the rate of expansion of the universe at a particular time, this will tell us the universe's geometry - a property known as curvature - at different epochs. The standard, homogeneous cosmological model predicts that this curvature evolves smoothly, making it an easy matter to calculate the geometry of space today from measurements taken at any distance.

This results in a simple consistency check: take measurements at two or more distances, and see what value for today's curvature pops out in each case. If the values do not agree, something is wrong with the homogeneous model. "If the measurements at different distances imply different curvatures today, then the assumption on which the standard cosmological model is based - the Copernican principle - is wrong," says Ellis.

Ellis himself, together with Clarkson and Jean-Philippe Uzan of the Institute of Astrophysics in Paris, France, has developed a variation on this type of consistency check. It involves taking measurements over about a decade of how fast single cosmological objects, such as quasars, are moving away from us with the universe's expansion, and how that motion changes at different distances - and therefore at different epochs. The results will tell us how the universe's rate of expansion has changed over time, and this can be checked against the predictions of the homogeneous cosmological model (Physical Review Letters, vol 100, p 191303).

The observational sensitivity required to record such tiny motions - the change in distance over 10 years is typically less than a billionth of the distance of the objects from us - is currently beyond astronomers' capabilities. But it should become feasible, the researchers suggest, with a new generation of ultra-sensitive telescopes, such as the Extremely Large Telescope planned by the European Southern Observatory.

The barrage of new work on testing the Copernican principle has certainly stimulated interest, with a flurry of papers in the past year proposing refinements or completely different tests. So what would it actually mean if, against the expectation of Caldwell and the majority of cosmologists, the outcome were that the Copernican principle is wrong?

It would certainly require a seismic reassessment of what we know about the universe. Our big bang model is particularly simple, characterised by universal quantities such as matter and energy densities and the rates of expansion and acceleration, whether negative or positive. If the Copernican principle fails, all that goes out of the window too. In that case, quantities we measure in our own rather special corner of the universe will turn out to have only parochial significance, with no deeper universal meaning. We would no longer be sure what, if anything, we can conclude about the wider universe, its origin, evolution and fate. Cosmology would be back at the drawing board.

If we are in a void, answering how we came to be in such a privileged spot in the universe would be even trickier. But regardless of how uncomfortable the philosophical implications might be, for Ellis it is a matter of scientific principle to test cherished but untested assumptions. "Whatever our theoretical predilections, they will in the end have to give way to the observational evidence," he says.

The Copernican principle might survive the tests, leaving us with the known unknown of dark energy. Or it might fall, leaving us with the unknown unknown of an entirely new cosmological model. Either way, cosmologists will still have plenty of explaining to do.

  12/11/2008. New Scientist Magazine.