Hubble Presses Toward Cosmic Dawn

Highest-redshift galaxies
The Hubble Ultra Deep Field is the deepest picture of the sky ever taken, showing objects to about magnitude 30.7. In it, astronomers have identified dozens of tiny, irregular, very red galaxies that have redshifts of 5 to 7, such as those circled in this small part of the image. These are among the earliest star-forming galaxies in the universe.
Courtesy NASA / ESA / Rogier Windhorst / Haojing Yan.
In March, NASA unveiled the Hubble Space Telescope's Ultra Deep Field (UDF), an 11.3-day exposure that represents humanity's deepest optical look into the universe. Althought it’s smattered with beautiful spiral and elliptical galaxies lying a few billion light-years from Earth, several teams of astronomers have focused on the faintest, farthest objects in the UDF — tiny red smudges that are barely discernible. While not exactly eye candy, these smudges were primarily responsible for one of the universe's most critical transitions: reionization.

The story begins moments after the Big Bang, when the universe was so hot that matter existed in an ionized state: a sea of atomic nuclei and free electrons. At an age of about 380,000 years the universe finally cooled enough so electrons could combine with nuclei to form hydrogen and helium atoms; the universe became neutral. But distant quasars seen in the Sloan Digital Sky Survey confirm that all intergalactic gas was ionized again by a cosmic age of 1 billion years, and it remains so to this day. What kind of objects reionized the gas, and when did they do the deed?

Highest-redshift galaxies
The Hubble Ultra Deep Field is the deepest picture of the sky ever taken, showing objects to about magnitude 30.7. In it, astronomers have identified dozens of tiny, irregular, very red galaxies that have redshifts of 5 to 7, such as those circled in this small part of the image. These are among the earliest star-forming galaxies in the universe.
Courtesy NASA / ESA / R. Windhorst / H. Yan.

On September 23rd, five teams reported UDF results at a workshop held at the Space Telescope Science Institute in Baltimore, Maryland. All the teams agreed that the red smudges represent small galaxies in the very distant universe bursting with the light of young stars. But they do not agree on how much these objects contributed to reionization.

The teams collectively identified roughly 100 galaxies in the UDF lying at redshifts of about 6, which corresponds to a cosmic age of 1 billion years. Follow-up spectroscopy with HST, the Keck Telescope, and the Gemini Telescope confirmed the redshifts. The galaxies appear to be roughly one-tenth the size of the Milky Way, but with star-forming rates 5 to 10 times higher. Many of the dwarfs appear close together, meaning they were destined to merge and form larger galaxies like our own.

A team led by Andrew Bunker (University of Exeter, England) concludes that these early galaxies were not forming stars fast enough to reionize the universe. "We did not see as much star formation as we expected," said Bunker. "There just aren't enough photons at that distance." He and his colleagues suggest that if these early galaxies created higher-mass stars than galaxies do today, or if there was an even earlier generation of star formation, that could explain why the universe was reionized by redshift 6. But Bunker added, "What we see and what we are sure of cannot have reionized the universe."

Highest-redshift galaxies
The Hubble Ultra Deep Field is the deepest picture of the sky ever taken, showing objects to about magnitude 30.7. In it, astronomers have identified dozens of tiny, irregular, very red galaxies that have redshifts of 5 to 7, such as those circled in this small part of the image. These are among the earliest star-forming galaxies in the universe.
Courtesy NASA / ESA / R. Windhorst / H. Yan.

On the other hand, Massimo Stiavelli (Space Telescope Science Institute) and his colleagues conclude that the population of dwarf galaxies seen in the UDF could have reionized the universe without much help from other objects. They made a reasonable assumption that these early stars formed from clouds that lacked elements heavier than hydrogen and helium. Such stars would have been more massive and burned hotter than stars formed with even small amounts of heavy elements (known as metals), so they produced more ultraviolet light, which ionizes gas very efficiently. "We think the galaxies that Andy Bunker found can do the job of reionizing the universe, if their stars are metal poor," said Stiavelli.

Haojing Yan (Caltech) and Rogier Windhorst (Arizona State University) took a different approach. They noted that the number of distant dwarf galaxies increases right up to the UDF's threshold of detectability, so these objects must represent the tip of the iceberg of a much larger population. "We need to correct for the incompleteness of the survey, or we will get a distorted picture of the distribution," says Yan. Certainly, Yan and Windhorst argue, there are even fainter objects at these distances that aren't seen in the UDF. By estimating the number of unseen objects, they find that the total output of light from dwarf galaxies should have been enough to reionize the universe. "We are confident these objects are real," said Yan. "We don't need new classes of reionization sources."

Faint galaxies
Three galaxies that do not appear in the ACS visible and near-infrared Hubble Ultra Deep Field (left column) show up in the NICMOS infrared UDF (right column). The galaxies appear in the infrared but not at shorter wavelengths because cosmic expansion redshifts their light output into the infrared. The letter z in each image represents the galaxy's redshift. NICMOS found a total of 6 galaxies with redshifts between 6.5 and 7.9. The fact that more galaxies aren't seen by NICMOS strongly suggests that these objects represent the initial burst of galaxy formation, about 700 million years after the Big Bang.
Courtesy Rychard Bouwens, Garth Illingworth (University of California, Santa Cruz), Rodger Thompson (University of Arizona), and NASA/ESA.

Garth Illingworth (University of California, Santa Cruz) and his colleagues scoured the two different versions of the UDF: one taken in visible and near-infrared light with Hubble’s Advanced Camera for Surveys (ACS), and one taken in the infrared with the Near-Infrared Camera and Multi-Object Spectrometer (NICMOS). They found several candidate objects in the NICMOS image that do not appear in the ACS image. Illingworth thinks these are dwarf galaxies at redshifts of 7 to 8, meaning we see them as they existed only 650 to 800 million years after the Big Bang. These galaxies don't appear in the ACS deep field because cosmic expansion has redshifted their starlight out of the visible part of the spectrum. Presumably, dwarf galaxies formed relatively quickly after the Big Bang, perhaps when it was half a billion years old (redshift 10). They would have commenced reionization almost as soon as they started, and they finished their job a half billion years later.

The final team, led by James Rhoads and Sangeeta Malholtra (Space Telescope Science Institute), identified a pronounced cluster of dwarf galaxies at a redshift of 5.9. They noticed the initial clustering in the UDF, but they found even more cluster members with the 4.0-meter Blanco Telescope at the Cerro Tololo Inter-American Observatory in Chile. The discovery shows that galaxies began to cluster in sheets and voids very early in cosmic history. "This is the most distant structure of this type yet seen," says Rhoads. He and Malholtra argue that reionization must have happened at different times in different places, as regions with a high density of galaxies were reionized more quickly than low-density regions. "Reionization proceeds in a patchy fashion," said Rhoads. "Regions with more light burned off the fog earlier."

Despite the lack of consensus on whether these dwarf galaxies alone could reionize the universe, independent commentator Richard Ellis (Caltech) pointed to the remarkable agreement on what has been seen. Theorists, Ellis noted, have speculated about this era of cosmic dawn for decades. "We're moving out of era of theory into an era of observation," he said. "This is a milestone. The sequence of deep fields from Hubble and other telescopes have changed the way we do astronomy. We have many unsolved issues, but we have a clear path forward."

All the workshop participants stressed that current observations push right to the limits of current instrumentation. To understand the early universe and the amount various sources contributed to reionization, it will be critical to fly at least one more servicing mission to Hubble to install the new Wide Field Camera 3, an advanced infrared instrument that has already been built and that should see even farther back in space and time than ACS and NICMOS. To penetrate all the way back to the era when galaxies first began to assemble will require the completion of Hubble's successor: the James Webb Space Telescope, which is currently scheduled to launch in 2011.