If you could see the universe as it was about 13.8 billion years ago, it would look like a flame. Back then it was just a hot ionized fog—a plasma—still glowing from its birth in the big bang at the dawn of time. But when it was still in its infancy, a youthful 370,000 years old, everything changed. Slowly cooling as it expanded, the universe grew chilly enough for electrons in the plasma to combine with protons, forming hydrogen gas. When the opaque plasma transformed into see-through gas, the plasma's last, fiery light was suddenly freed to flash through a newly transparent universe.

That ancient, relic light washes over us even now, diminished by the intervening eons to a faint all-sky microwave glow: the cosmic microwave background (CMB). This light gives cosmologists their earliest possible glimpse of the infant universe, providing crucial evidence for its origins, age and composition. But besides being the ultimate cosmic baby picture, the CMB also provides snapshots of the universe as it has grown up. Passing through billions of light-years of expanding space on its way to us, the CMB has been subtly altered by interactions with coalescing cosmic structures. It encodes the story of how a nearly featureless soup of matter and radiation came to be our orderly cosmos of galaxies, stars and planets.

Now, cosmologists using CMB maps from the European Space Agency's Planck satellite have pinned down the timing for one key event in that story—the end of the universe's “Dark Ages,” the epoch that lies between the CMB’s creation and the formation of the first stars. According to Planck's data, gathered between 2009 and 2013, starlight began flooding the universe 560 million years after the big bang. This date is 140 million years later than the previous best estimate of 420 million years post–big bang, which came in 2006 using CMB data from the Wilkinson Microwave Anisotropy Probe (WMAP). The later arrival indicated by Planck is good news for observers hoping to see the light from the first stars with next-generation instruments that will soon come online, like the James Webb Space Telescope and the Square Kilometer Array. It's also good news for theoretical cosmologists, who had struggled to reconcile WMAP's results with other conflicting estimates for the Dark Ages' end.

Shedding light on the Dark Ages
When the first stars formed, their intense ultraviolet light “reionized” the universe, stripping electrons from much of the hydrogen. The resulting electron-rich plasma scattered the CMB's photons as they passed through. This scattering is imprinted in the CMB's polarization (the direction in which waves of light oscillate as they travel). The earlier the first stars formed, the more scattering electrons they would produce, and the stronger the resulting polarization signal would be in the CMB. It was this polarization that WMAP and, later, Planck, attempted to measure.

“The WMAP analysis placing the reionization at 420 million years after the big bang was a real puzzle,” says George Efstathiou, a University of Cambridge cosmologist and a leader of the Planck Collaboration. “Because there just weren't enough stars around back then to reionize the universe so quickly.” After 2006's WMAP announcement cosmologists pushed the Hubble Space Telescope to its observational limits, conducting several deep surveys in search of the early star-forming galaxies required to support the result. The surveys simply didn't find many luminous objects at sufficiently early times to support WMAP's timing—there didn't seem to be enough starlight during that era to reionize the universe.

Theorists, however, already had a handful of possible explanations for how the universe could have become so illuminated so fast. Perhaps, some thought, there were hordes of small galaxies too dim for Hubble to see. Or maybe, others said, the first stars were strange, short-lived and supermassive giants, far brighter and hundreds or even a thousand times more massive than our sun. These ideas have not been disproved, but thanks to Planck's revised timeline, Efstathiou says, “we now don't need anything other than what we already see.”

“A difference of 140 million years may not seem like much, but it is sufficient to explain the reionization without an early generation of stars or other exotic processes,” says Marco Bersanelli, a Planck collaborator at the University of Milan. “The Planck results indicate a nice simplification of our understanding of this crucial epoch of cosmic history.”

Princeton University's David Spergel, a prominent former member of the WMAP team, agrees. “This really helps reinforce the standard cosmological model, and that makes people's lives easier,” Spergel says. Planck's latest results, he notes, are actually broadly consistent with WMAP's earlier data, although they show that WMAP's measurements were contaminated by a polarization signal from dust in our own Milky Way. When the galactic-dust data from Planck's more sensitive instrumentation is used to correct the WMAP findings, the resulting estimate for reionization also comes out to about 560 million years after the big bang. Additionally, the Planck collaboration performed one more check, using a technique called gravitational lensing to scour the CMB for evidence of how clumpy the universe was in its early life. This technique can yield an entirely independent estimate for the timing of reionization. “Lensing has nothing to do with polarization, and that's important,” Efstathiou says. “Yet it gives the same answer: 560 million years.”

A brighter future
Other independent checks are already streaming in, including a new analysis combining the latest constraints from Planck with the previously puzzling Hubble Space Telescope surveys. “Our paper shows that early star-forming galaxies likely supplied enough ionizing radiation to explain the timing and magnitude of reionization inferred from the Planck data,” says lead author Brant Robertson of the University of Arizona. The results not only validate Planck, but also hint at what NASA's James Webb Space Telescope, built to see the first stars and galaxies, may witness after its scheduled launch in 2018. The joint Planck–Hubble data, Robertson says, suggest that Webb could easily see on the order of five galaxies less than half a billion years old “right off the bat in its first year of observations.”

“By studying the first luminous objects formed, Webb and instruments like the Square Kilometer Array hope to see and understand how the transition occurred from a universe with nothing to one with stars and galaxies,” Efstathiou says. “Planck's results suggest figuring this out will be easier. The later the first luminous objects formed, the better, the more we have a chance of seeing them.”

Beyond the question of when the universe's first stars formed, the Planck data may still solve other mysteries—and reveal new ones. According to Bersanelli, a late end to the Dark Ages “favors a simple cold dark matter scenario” where massive, slow-moving and weakly interacting particles make up most of the elusive, unknown substance that overwhelmingly outnumbers visible matter in the universe. Other alternative explanations, such as “hot” scenarios where dark matter particles whiz around at substantial fractions of the speed of light, are more difficult to reconcile with the corpus of data on the chronology of cosmic structure formation.

Planck's latest data also seems to weaken the cases for the simplest models for cosmological inflation, a postulated acceleration in the expansion of the primordial universe in the first instants after the big bang. “These models aren't quite ruled out by Planck, but it's not looking good,” Spergel says. The most intriguing part of the new Planck data, he notes, is what it suggests about the “clumpiness” of matter in the early universe. Measuring this can help constrain models of dark energy, the mysterious force that seems to be driving the universe's present-day accelerating expansion. “As inferred by Planck, the universe looks lumpier back then, by a couple percent, than what we observe looking at galactic clusters. There are three possible explanations: Maybe there's something wrong with the CMB data.... Or maybe there's something wrong with how we're interpreting the counts of galactic clusters. The final possibility, which would be the most exciting, is that we need new physics!”


Origami Observatory: Behind the Scenes with the Webb Space Telescope [Preview] [https://www.scientificamerican.com/article/origami-observatory/]

The First Stars in the Universe [preview] [https://www.scientificamerican.com/article/the-first-stars-in-the-universe-2004-09/]

Continental Telescope Array Could Usher Astronomy Revolution in Africa [https://www.scientificamerican.com/article/continental-telescope-array-could-usher-astronomy-revolution-in-africa/]