FOUR years ago disaster struck the Earth, borne on a beam of light. On its 12-billion-year journey from a distant quasar, the light had passed through interstellar clouds of metals such as iron, nickel and chromium. When astronomer John Webb and his team at the University of New South Wales in Sydney analysed the light, they found these atoms had absorbed some of its photons. But, according to the known laws of physics, they had absorbed the wrong ones. Something was amiss.
This single observation rips apart our most cherished theories of how the Universe works, according to physicists Thomas Banks and Michael Douglas of Rutgers University, New Jersey, and Michael Dine of the University of Santa Cruz. If the astronomers really saw what they thought they saw, then our best explanation for how matter and forces interact – the standard model – is proved woefully inadequate. M-theory, the most successful incarnation of the idea that all matter is made up of vibrating strings, can’t cope with the observation either. Banks, Dine and Douglas make this claim in a paper to be published in a forthcoming issue of Physical Review Letters. And they don’t make it lightly, since all three have been personally involved in developing M-theory.
The implications of Webb’s data are huge. Incorporating the astronomers’ finding into M-theory or the standard model would mean changing the theories so much they no longer explain accepted concepts such as inflation, which describes how the Universe expanded rapidly after the big bang. It would mean twisting theories so far that the long-hunted Higgs particle should already have turned up in our particle accelerators. In other words, with our current understanding of the Universe, it can’t be done.
Advertisement
How can a flash of light wreak such havoc? Because it suggests that one of the pillars of modern physics – the fine structure constant, or alpha – is not constant at all. Alpha is a “coupling constant”: it dictates how photons interact with particles such as electrons, describing, for instance, which photons will be absorbed by the electrons in a cloud of metal atoms. Today, alpha is around 1/147. But in that original observation of light from the quasar, Webb found that alpha was slightly smaller by about a millionth of its present value. Twelve billion years ago, iron and magnesium absorbed photons of different energies than the ones they absorb today.
In 1999, when Webb first announced that alpha might have changed, physicists were more sceptical than worried. The first question on everyone’s lips was: “Is it experimental error?” Banks was no exception. “My first reaction to this is that the observations should be rechecked very carefully,” he says.
That’s exactly what Webb and his team have been doing. They spotted the anomalies in the light because of an innovative analysis technique first used in the 1999 results. This improved the sensitivity of the Keck telescope in Hawaii, where all of their observations have been made, by a factor of 10. They have now collected and analysed light from more than a hundred quasars, and so far every observation backs up the shift in alpha. Webb intends to publish the new data later this year. “These results are going to cause quite a fuss when they come out,” he says.
He is anything but complacent, however. His team is checking for a systematic error introduced by their equipment. “We’re doing our best to find out what we could have done wrong, and what that systematic error could be,” he says. So far they’ve found no error that might explain their result (Monthly Notices of the Royal Astronomical Society, vol 327, p 1208, 1223, 1237 and 1244).
As a final check they have just begun an analysis of quasar light caught by the European Southern Observatory’s Very Large Telescope in the Atacama Desert in northern Chile. This involves a different instrument, different observers and a different technique for handling the data, so these results should expose any systematic errors. “My suspicion is that, if it’s not a systematic effect, it’s varying alpha,” says Webb.
John Barrow of the University of Cambridge, who has been working with Webb, is more bullish. There would have to be an unimaginable sequence of coincidences to get such a consistent error, he says. The agreement in the value of alpha measured from so many quasars is “uncanny”, says Barrow, and a varying alpha is now pretty much undeniable. “Statistically, it’s of vastly higher significance than you need to detect an elementary particle at CERN.”
And if Webb’s measurements are correct, every area of physics touched by alpha will need to be revisited. Alpha is a conglomerate of other constants, given by the formula 2πe2/hc, where h is Planck’s constant, e is the charge of an electron and c is the speed of light. At first, the notion that alpha could vary set researchers wondering whether they could accommodate the idea. For example, some have suggested that the electron’s charge or the speed of light has varied in the past. It’s a hugely controversial idea, but it could explain the observed data and solve some other puzzles of cosmology (see “All change please”).
For the theorists, though, simplistic solutions like this just won’t wash. The trouble is that the value of alpha governs many other processes, such as the strength of the weak nuclear force that affects how the Sun burns and how radioactive beta decay occurs. It governs the “inflation” of the Universe after the big bang. Thus alpha provides a way to predict the rate at which these processes happen, and all the predictions based on today’s alpha are supported by experiments.
Take the fluctuations in “empty” space – the quantum vacuum. In a vacuum, pairs of particles and antiparticles are constantly springing into existence and then being annihilated almost immediately. But the process of creating and destroying an electron-positron pair also creates and destroys a photon. “And so alpha controls the strength of these processes,” says Douglas. Physicists can check their calculations for how much energy these processes need, and hence the rate at which they occur, against experimental evidence. And so far, the experiments fit with calculations using today’s value of alpha.
But change alpha to be consistent with Webb’s observations, or invoke an alpha that changes over time, and physics simply cannot cope. Physicists’ calculations would no longer agree with what they measure in experiments. The vacuum energy is particularly important because it is directly related to the cosmological constant, lambda, which describes the accelerating expansion of the Universe. Lambda is extremely sensitive to the strength of alpha and any variation would make physicists’ theoretical early Universe blow up ridiculously fast. “All of conventional cosmology would be affected in a way that is grossly inconsistent with observation,” says Banks.
It wouldn’t be the first time physicists have been forced to indulge in a little creative thinking to make their ideas work, however. Theoretical physics predicts a value for lambda that is 10120 times higher than the observed value. But, the scientists argue, this simply means there’s an as yet undiscovered physical mechanism that reduces it to the observed value. When they incorporate this “fudge factor” into their models, they can reduce the vacuum energy to fit with observations without affecting anything else.
But that doesn’t deal with a varying alpha, because change alpha and you somehow have to juggle changes in several other parameters in the theoretical models – including the fudge factor that makes the cosmological constant fit so neatly with observations.
So if Webb’s data and the theorists’ prognosis hold up, there’s only one possible outcome: we can wave goodbye to our “understanding” of the Universe. “If these observations are confirmed, one will have to invent some very exotic physics to explain them,” Banks says.
It’s not all gloom though. A varying alpha brings some benefits. It could solve cosmology’s “horizon problem”, for example. Physicists have a hard time explaining why far-flung parts of the Universe are all at roughly the same temperature. It implies that these areas were once close enough for energy to pass between them, evening out their temperatures. Yet models of the early Universe prevent this from happening.
One way round this is if the speed of light was higher in the infant Universe than it is today, allowing energy in the form of light to pass between these areas (Â鶹´«Ă˝, 24 July 1999, p 28). And if the speed of light has decreased over the lifetime of the Universe, that would explain why alpha has increased.
This isn’t the only long-standing puzzle that changes in alpha would resolve. Paul Langacker, Gino Segrè and Matthew Strassler at the University of Pennsylvania have worked out how shifts in alpha could change the way elements such as helium formed in the early Universe. As the Universe cooled after the big bang, a time came when there was no longer enough energy for the weak nuclear force to change protons into neutrons and vice versa. From then onwards, the relative abundance of protons and neutrons was fixed.
Astronomers calculate that this event set a ceiling on the number of helium nuclei, each of which contains two protons and two neutrons, that could be created. But the amount of helium floating around just after the big bang was vastly more than such theories predict – unless, that is, the strength of the weak nuclear force has changed. The value of alpha determines the effects of the weak interaction. If alpha changed, so would the relative abundance of helium and hydrogen created after the big bang. In fact, says Langacker, we might be able to use readings of how much helium was formed in the early Universe to work backwards and find the value of alpha at that time.
Langacker, himself a string theorist, doesn’t believe the varying alpha is a problem for all our models of the Universe. He feels that Banks, Dine and Douglas are overreacting: string theories such as M-theory do allow physical “constants”, such as alpha to vary over time, he points out. Examining the way these constants vary may even give us a clue to the more fundamental physics behind them, he argues.
But Douglas disagrees: he believes variations of the kind Webb’s data suggest would be a serious problem. While M-theory can describe varying coupling constants, this is one of its main problems when tested against real world observations, Douglas says. In many M-theory scenarios, varying coupling constants directly contradict experimental observations. The business of developing ideas to fix this problem is a major subject of research. “Webb’s observations – if true, and if they do not have some other explanation besides varying coupling constants – have dire consequences for these ideas,” Douglas says.
But he thinks that a varying alpha could ultimately prove to be a good thing. In the past, such crises have helped to refine physics. “It might well be the key to a real understanding of these issues,” Douglas says. Banks is less optimistic: if alpha is varying, he doesn’t see how physics can get us past the problem. “If the observations stand up, and the only explanation for them is variation of alpha, I think that it means our current theoretical understanding is seriously flawed,” he says.
So, if physicists are as objective as they claim to be, and accept the observations as valid, we could well be living with some new “facts” by the end of the year: that a number of our fundamental constants are not constant after all; that M-theory fell at its first experimental hurdle; and that the standard model is a shoddy, incoherent explanation for the subatomic world. And all because of some aberrant light from across the cosmos.
“If there is some independent confirmation of these results, this issue will move to the forefront of physics,” says Dine. “Some sort of surprising – and currently unknown – physical mechanism or principle must be at work.”
We’ll soon know if he’s right.
All change please
John Barrow likes the idea of a varying alpha. The Cambridge University cosmologist certainly has no time for the bleatings of theoretical physicists who say the observations must be wrong because they don’t fit the models. “Their story would predict a cosmological constant 10120 times bigger than what we see, so I’m a little wary of those sorts of arguments,” he says. The way to resolve the debate is through experiment, he argues, not by someone saying: “Well, it’s awfully difficult to fit in.”
Barrow and his colleagues Joao Magueijo and Havard Sandvik of Imperial College in London have built a model of the cosmos in which the speed of light or the electron charge can change to make alpha vary. They’re very pleased with the result, published earlier this year (Physical Review Letters, vol 88, p 31302). “It matches the astronomical results rather naturally,” Barrow says.
Their model produces a varying alpha in only one epoch of the evolution of the Universe. This, they say, fits with the data from the Oklo natural reactor in Gabon (see “Down to Earth”) – which limits recent variations in alpha to being vanishingly small – and with quasar data that suggests comparatively large variations in alpha much longer ago.
But simply tweaking the electron charge or light speed is unsatisfactory, counters Xavier Calmet, a theorist at Ludwig-Maximilians University in Munich. “The problem is, if you have variation of the speed of light, you need some special kind of physics,” he says. “What would cause such a variation?”
Down to Earth
THE best earthbound way to measure the fine structure constant, alpha, depends on a rare natural phenomenon in West Africa. Two billion years ago, under what is now Oklo in Gabon, a deposit of uranium ore heated up like a nuclear reactor. There wasn’t enough of the radioactive metal to cause an explosion, but the “reactor” left behind a unique record of fission products. The abundance of those products – determined by alpha-dependent processes such as neutron capture and beta decay – provides a way to measure how big alpha was all those years ago.
In 1996, Freeman Dyson and Thibault Damour of the Princeton Institute for Advanced Studies in New Jersey concluded that if alpha had changed at all, it was by a billionth of its present value. In other words, a vanishingly small amount.
This appears to conflict with Webb’s measurement of alpha from quasar light, and has been widely used to suggest that the astronomers must be wrong. But interpretations of the Oklo data are based on assumptions that could well be wrong. For a start, no one can be sure that alpha has always changed at a constant rate, so there’s no need for a conflict with the astronomers. And then there’s the possibility that a changing alpha may have shifted other factors, such as the neutron/proton mass ratio, the mass of a subatomic particle called a pion and the strength of the strong force that holds protons and neutrons together in atomic nuclei. These changes could somehow skew the results from Oklo.
For this reason, the constraints on a varying alpha imposed by Oklo “cannot be taken seriously”, according to a paper by Xavier Calmet of Ludwig-Maximilians University in Munich, Germany, and Harald Fritzsch of Stanford University. Only a new analysis that accounts for all the possibilities will solve the apparent conflict between data from Oklo and that from quasars.
- Further reading: “The cosmological evolution of the nucleon mass and the electroweak coupling constants” by Xavier Calmet and Harald Fritzsch ()
- Further reading: “Time-varying alpha and particle physics” by Thomas Banks, Michael Dine and Michael Douglas () “Implications of gauge unification for time variation of the fine structure constant” by Paul Langacker, Gino Segrè and Matthew Strassler ()