IN THE past few years, a consensus has emerged about the kind of Universe we live in. It is thought that some kind of repulsive force, which has come to be known as dark energy, is pushing the Universe into an accelerating expansion.
Is all well with this impressive and, for cosmologists, unusual consensus? I鈥檓 not so sure. In a paper published last month in Monthly Notices of the Royal Astronomical Society, I argue that the evidence is not as solid as it seems. There are grounds for doubt that dark energy exists at all.
In 1998, Science magazine rated universal acceleration as the most significant discovery of the year. The two teams involved鈥攖he Supernova Cosmology Project (SCP) led by Saul Perlmutter of the University of California at Berkeley, and the High-Z Supernova Team (HZT) led by Brian Schmidt of Mount Stromlo Observatory, Australia鈥攈ad discovered several distant examples of a particular kind of supernova.
Advertisement
Called a type Ia, it is the explosion of a white dwarf star in a binary star system. Material from a companion red giant star is dumped on the white dwarf until the smaller star reaches a precise mass limit. At that point the white dwarf can no longer support its own weight, and burns its nuclear fuel so suddenly that it explodes.
These explosions always release roughly the same amount of energy, and studies of relatively nearby type Ia supernovae have shown that they reach almost the same peak brightness in every case. So type Ia supernovae can be used as 鈥渟tandard candles鈥� to determine their true distance.
We know how fast the Universe is expanding today. If the expansion is accelerating, then it must have been slower in the past. This means the 鈥渓ook-back鈥� time to a distant galaxy will be greater than expected, which means that its light has had longer to spread out and it looks fainter than if the Universe was expanding at a steady rate. High-red-shift supernovae are found to be about 20 per cent fainter than expected. Both teams interpret this as evidence for acceleration.
This came as a huge surprise, as everyone had assumed that the Universe鈥檚 expansion was slowing down, not speeding up. After all, the only force known to be involved on these huge scales is gravity, which will always tend to pull galaxies together and so decelerate the expansion of the Universe. Acceleration implies that there is some new force, some kind of universal repulsion.
In fact, the idea is not new. In 1917, Einstein tried to make a model of the Universe within the framework of his general theory of relativity. He introduced an additional term in his equations, the cosmological constant, usually labelled lambda. He did this to make his model of the Universe static, with a repulsive lambda-force balancing the attractive force of gravity.
When the expansion of the Universe was discovered a decade later by Edwin Hubble, Einstein regretted modifying his equations in this way and referred to it as mein gr枚sster Fehler鈥攎y greatest mistake. But a repulsive force is just what鈥檚 needed to explain the supernova data, so lambda seems to be back.
The usual physical interpretation of lambda is that it is the energy density of empty space. According to particle physics, a vacuum seethes with virtual particles that come into existence for an instant and are then annihilated. Unfortunately, the calculations say that this energy density is huge鈥攁bout 10120 times as great as the value required by the supernova teams. Some as yet unknown mechanism in particle physics may cancel out most of this energy, but I find it easier to believe that such a mechanism would reduce it to exactly zero rather than the very low value that seems to be observed. That is one aesthetic reason to be sceptical of the supernova results.
But my real objection is that I don鈥檛 think the observational evidence is compelling. The first problem is that dust could be shrouding the explosions, making them look fainter and deceiving us into thinking that they are further back in time. Both the SCP team and the HZT team conclude that the effects of dust are small, but I disagree. A team led by Mark Phillips of Las Campanas Observatory, Chile, has found that for nearby supernovae, the explosions are dimmed by about 25 per cent on average. But the techniques used by the SCP and HZT teams assume a lower value than this.
Moreover, there are reasons for believing that at high red shift, dust would cause still greater dimming. Because we are looking back in time, distant galaxies are younger as we see them than nearby ones. And younger galaxies tend to hold more gas and dust.
The second problem is to do with those nearby type Ia supernovae that were used to show they all have the same peak brightness. Quite a few of them were not observed until after they had reached their maximum and started to decline. The teams then extrapolated back to the maximum. Although this practice should be valid on average, it has the effect of artificially reducing the scatter of the data and hence underestimating the statistical uncertainty of the final result.
When I reanalyse the supernova data and correct for these two effects, including a modest correction for the increased dimming in younger galaxies, the evidence for acceleration disappears.
Many people I present this argument to say that their belief in dark energy doesn鈥檛 depend on the supernova evidence alone: there are three reasons for believing in dark energy. So even if we completely demolish the supernova argument, the edifice of dark energy is still standing, if only on two legs. However, both these remaining supports are built on questionable foundations.
One argument is that vacuum energy explains the inconvenient age of the Universe. The age of certain star clusters in our Galaxy can be measured from the colours and brightnesses of the stars, and various studies imply that the Galaxy itself is between 12 and 13 billion years old. But according to the latest observations from the Hubble Space Telescope, the Universe is expanding quite rapidly. This affects how the age of the Universe is gauged. If you run the expansion backwards you can see how long it takes until everything was at the single point of the big bang. The faster the expansion, the shorter the time. If only gravity acts to slow the expansion, it must have been going even faster than that in the past, and so the time taken is shorter still. In that case, and if the Universe has a flat geometry (see below), the Hubble measurement of expansion means the Universe is about 9 billion years old, much too young to contain the oldest stars in our Galaxy. But with cosmic acceleration, the expansion would have been slower in the past, and the age problem is fixed.
However, the measurement of the current expansion rate depends on an assumed distance to the Large Magellanic Cloud, and the typical uncertainty in measurements of this distance is about 10 per cent either way. It could be out by as much as 20 per cent, and so there鈥檚 a chance that the expansion age is more like 12 billion years even in a flat, matter-dominated universe. A universe with zero vacuum energy is not ruled out.
A more impressive line of evidence is based on the cosmic microwave background radiation. This is a relic of the hot early Universe, about 300,000 years after the big bang. Over the past few years, maps of this radiation have been made using balloon-borne and ground-based microwave telescopes, notably by the MAXIMA, BOOMERANG and DASI teams. These maps show ripples in the background radiation, a sign that the soup of matter and radiation that filled the early Universe was undergoing oscillations. The whole Universe was resonating to these giant sound waves, and it鈥檚 relatively easy to work out what wavelengths they had. So we can compare that size with the observed angular scale to test the geometry of the Universe: if space is curved it will act like a lens, and either magnify or shrink the images of these oscillations.
In fact, there doesn鈥檛 seem to be any distortion, so the Universe is probably flat. And that is taken as evidence for dark energy, because general relativity says that for space to be flat there has to be a particular average energy density throughout the Universe. One source of that energy density is matter, but there doesn鈥檛 seem to be enough matter around, even when invisible 鈥渄ark matter鈥� is taken into account. The most obvious candidate to make up the difference is some kind of vacuum energy.
But there is a small loophole in this argument. The total amount of matter is measured by looking at how galaxies cluster together. Galaxies are grouped into clusters, and those clusters of galaxies are grouped into superclusters, but here is a characteristic scale above which the clustering rapidly falls off. This scale is determined by a tug-of-war in the early Universe between slow-moving stuff (heavy particles of matter), and fast-moving stuff鈥攔adiation and other particles moving close to the speed of light. More matter and the scale decreases, so there are fewer big superclusters; less matter and it increases, so there are more superclusters.
But the balance with radiation depends on how many different kinds of relativistic particle there are. We think we know the answer: photons, plus three kinds of neutrino (electron, muon and tau neutrinos). But it鈥檚 possible that there were other kinds of neutrino or other relativistic particles at that early time that have subsequently decayed. If so, we have got the estimate of matter density wrong, and there could be enough matter to make space flat after all.
With such a range of evidence in favour of dark energy, it might seem churlish to raise these doubts. But that鈥檚 what science is about鈥攓uestioning the evidence and testing alternative ideas. And if dark energy does not exist, we have a much simpler and more elegant model for the cosmos.
The Universe would be spatially flat and would be close to the critical density, so that the expansion continues forever, but slows down as time goes on. Of course, we would still be left with the mystery of what makes up dark matter, but at least two mysteries would have become one.