Long after the big bang, the expansion of the universe has again begun to accelerate, as shown in this diagram.

NASA/WMAP Science Team

Is dark energy an illusion?

For the past 20 years, physicists have known that the expansion of the universe is accelerating, as if some bizarre “dark energy” is blowing up space like a balloon. In fact, cosmologists’ well-tested standard model assumes that 69% of the content of the universe is dark energy. However, there may be no need for the mysterious stuff, a team of theorists claims. Instead, the researchers argue, the universe’s acceleration could be driven by variations, or inhomogeneities, in its density. If so, then one of the biggest mysteries in physics could be explained away with nothing other than Albert Einstein’s familiar general theory of relativity. Other researchers are skeptical, however.

“If it’s right, somebody is going to have to take back Nobel prizes” awarded in 2011 for the discovery of the accelerating expansion of the universe, says Nick Kaiser, a cosmologist at the University of Hawaii in Honolulu. Tom Giblin, a computational cosmologist at Kenyon College in Gambier, Ohio, who has worked on a similar analysis, says, “I would love if inhomogeneities explained dark energy.” However, he says, “I don’t see any evidence from our simulations to expect it to be as big an effect as they see here.”

At issue is the way cosmologists calculate how the universe evolved over the past 13.8 billion years. Roughly speaking, they rely on two equations. One describes how matter coalesces into galaxies and clusters of galaxies. The other, known as the Friedmann–Lemaître–Robertson–Walker (FLRW) metric, comes out of Einstein’s theory of gravity, or general relativity, and scientists use it to calculate how much the universe has expanded at any time. At each step in time in a simulation, the cosmologists’ program uses the FLRW metric to calculate the “scale factor,” which specifies how much the universe has grown. The program then uses the scale factor as an input to calculate how the formation of galaxies and clusters advances in that step.

Strictly speaking, however, the FLRW equation applies to a smooth and homogeneous universe. So to calculate the scale factor at each step, cosmologists typically assume the universe is smooth and use its average density—determined from the simulation—as the FLRW metric’s input. That’s a bit dicey, because general relativity says that mass and energy warp spacetime. As a result, space should expand faster in emptier regions and slower in crowded ones, where the galaxies’ gravity pulls against the expansion. Thus, in principle, inhomogeneities in the universe can feed back through the dynamics and affect the universe’s expansion.

Gábor Rácz and László Dobos, astrophysicists at Eötvös Loránd University in Budapest, and their colleagues set out to capture that “backreaction.” They simulated a cube of space measuring 480 million light-years along each side. Instead of using the FLRW metric to calculate at each time step a single scale factor for the entire cube, they broke the cube into 1 million miniuniverses and then used the equation to calculate the scale factor in each of them. “We assume that every region of the universe determines its expansion rate itself,” Dobos says. The researchers then calculated the average of the many scale factors, which can differ from the scale factor calculated from the average density.

The team’s virtual universe evolved much as the real one has, with its expansion accelerating over the past few billion years. That happened even without adding space-stretching dark energy to the simulation, the researchers report in a paper in press at the Monthly Notices of the Royal Astronomical Society. The results suggest that it may be possible to explain away dark energy as an illusion, Dobos says.

Others are cautious. Giblin notes that the simulation he and his colleagues performed differs from the new one. The new work tracks the evolution of the universe to finer spatial scales, but involves certain assumptions and approximations, he says. In spite of the differences, Giblin says, his work suggests that backreaction would change the expansion rate of the universe by less than a percent, whereas the new simulation suggests an effect in excess of 20%.

Kaiser also expects the effects of inhomogeneity to be small. He notes that the best evidence for the accelerated expansion of the universe comes from measuring the distances and ages of stellar explosions known as type 1a supernovae in the relatively nearby universe. However, in the local universe, plain Newtonian gravity should work well enough. That suggests the difference in how the scale factor is determined in a relativistic theory shouldn’t exert a big effect. “If they’re right, there’s something very funny going on,” he says.

Still, experts say it’s reasonable to investigate backreaction. “I would say that this is now part of the mainstream in that people want to calculate the size of this effect,” says Thomas Buchert, a cosmologist at the University of Lyon in France, who pioneered the topic in the 1990s. Giblin notes “mainstream cosmology has done such a bad job of solving the dark energy problem that it will likely be some nonmainstream idea like this that does.” But, he adds, “I don’t know if this is the one.”