Imagine a universe with no stars, no galaxies, and no light: just a black brew of primordial gases immersed in a sea of invisible matter. Beginning a few hundred thousand years after the blinding flash of the big bang, the universe plunged into a darkness that lasted almost a half billion years. Then something happened that changed it all, something that led to the creation not just of stars and galaxies, but also of planets, people, begonias, and lizards. What could that something have been?
New clues to this puzzle—one of the most fundamental in cosmology—are pouring in from many directions. Theorists using supercomputer simulations have retraced the steps that produced the first stars and galaxies. Astronomers peering through giant new telescopes have journeyed back in time in search of the first galaxies. Researchers studying images from the Hubble Space Telescope have discovered the breathtaking diversity of the galaxies that surround us today—from giant pinwheels blazing with the blue light of newborn stars, to misshapen footballs glowing with the ruddy hue of stars born billions of years ago, to tattered galaxies trailing long streamers of stars torn out by collisions with intruder galaxies.
Less than a century ago astronomers knew only about our own galaxy, the Milky Way, which they believed held about 100 million stars. Then observers discovered that some of the fuzzy blobs in the sky weren't in our own galaxy, but were galaxies in their own right—collections of stars, gas, and dust bound together by gravity. Today we know that the Milky Way contains more than 100 billion stars and that there are some 100 billion galaxies in the universe, each harboring an enormous number of stars.
Our view of the universe is changing completely, says cosmologist Carlos Freak of the University of Durham in England, and it's largely because of our new understanding of galaxy formation: "It's no exaggeration to say that we're going through a period of change analogous to the Copernican revolution."
One of the new cosmologists, Tom Abel of Pennsylvania State University, thinks he has figured out how the first star was born. One afternoon last April he sat by a hotel pool in Cozumel, Mexico, oblivious to the squawking blackbirds and the whir of the poolside blender kept busy making piña coladas. He was staring intently at images on his laptop computer—images depicting how star formation could have happened. In a few minutes he would go back inside the hotel to share the images with his colleagues at one of the largest meetings ever devoted to the origin of galaxies.
The first star was born about 14 billion years ago, Abel believes, in a universe that was more mysterious but also far simpler than our own. Smaller and denser than today, the universe was pitch-black and contained mostly hydrogen and helium with a smattering of lithium. During the past few years, together with his colleagues Michael L. Norman of the University of California, San Diego, and Greg L. Bryan of Oxford University, Abel has created supercomputer simulations that show how stars were formed from these gases.
The first step, according to the simulations, was when gravity gathered gases into diffuse clouds. As the gases cooled, they coalesced at the center of each cloud into a clump no larger than our sun. The dump collapsed further, while surrounding gas piled on top of it. In this way it grew into a behemoth about 100 times the mass of the sun. Finally, several million years after the entire process began, the intense compression forged a full-fledged star—and there was light.
Elsewhere the same star-forming process had begun in other gas clouds that Abel refers to as microgalaxies—miniature, single-starred versions of today's galaxies. Soon beacons of light from massive stars permeated the darkness. These stars burned brightly and then fizzled after only a few million years, dying in titanic explosions called supernovae. During the brief time these first stars reigned, however, they wrought changes in the universe that had a profound effect on future galaxy formation. They heated surrounding gases and bombarded them with ultraviolet light. And when they exploded, the stars seeded the universe, and the next generation of stars, with the first supply of heavy elements, including the oxygen we breathe.
The explosive demise of these stars may have left behind dense cinders, the first black holes in the universe. Moreover, the supernova explosions may have been accompanied by flashes of energetic radiation known as gamma-ray bursts that are billions of times brighter than the sun. If so, some of the gamma-ray bursts that have already been detected may actually have come from the first stars.
"It would be the most wonderful thing," said Abel, "if we were so lucky that the first stars that formed were also the brightest."
Abel's presentation in Cozumel was a success. Scientists consider his simulations the most convincing scenario yet for how stars were born. The simulations are based on a mind-blowing concept: Some kind of mystery material, which can't be seen and has come to be known as dark matter, outweighs all the visible material in the universe by at least nine to one. Galaxies are merely bright flecks on a sea of dark matter. Without the extra tug provided by dark matter, astronomers say, there wouldn't be enough gravity to pull material into galaxy-size clumps or even form the first star.
The concept of dark matter has been around for decades, but cosmologists were slow to embrace it. That might have been because one of the first people to suggest it was a brilliant but abrasive genius named Fritz Zwicky, born in 1898. Zwicky's personality didn't encourage a fan club. He once called his colleagues at the Mount Wilson Observatory "spherical bastards," because, he said, they were bastards anyway you looked at them. In 1933 Zwicky turned his attention to a nearby cluster of galaxies, the Coma cluster, and realized it shouldn't exist. Individual galaxies in Coma were zipping around so fast that the gravity exerted by the visible parts of the cluster was too puny to keep Coma intact. But Zwicky had a solution. He proposed that all the visible material in the cluster was a mere fillip. The rest, which he could not see, he dubbed dark matter. No one wanted to believe "crazy Fritz" was right.
Decades later, resistance to Zwicky's ideas began to fade when astronomers found themselves invoking dark matter to explain a host of puzzles. In 1973 Princeton cosmologists Jim Peebles and Jerry Ostriker said the mystery material was necessary to keep spiral galaxies, including our own Milky Way, from falling apart. A few years later, Vera Rubin of the Carnegie Institution of Washington concluded that spiral galaxies she and her colleagues had examined had to be embedded in a halo of dark matter. That was the only way to explain, she said, why stars at the outer edge of the spiral galaxies moved no more slowly than stars at the core.
Dark matter, moreover, answered a key riddle of galaxy formation: how the universe changed from a smooth, hot soup of particles into a jumble of galaxies and galaxy clusters. There had to be some lumps in the first place. By itself, ordinary matter—protons, electrons, and neutrons—couldn't provide those lumps. There wasn't enough of it, and it couldn't begin clumping until the universe had cooled. Dark matter, by contrast, was plentiful and all but impervious to every force but gravity. It could coalesce almost immediately after the universe's birth, giving ordinary matter a foothold to form galaxies, even as cosmic expansion tried to pull them apart.
Evidence backing up the lumpy soup theory came in 1992, when a NASA satellite called the Cosmic Background Explorer detected tiny hot and cold spots in space. This supported the idea that the seeds of galaxy formation—the primordial lumps in the early universe created by dark matter—left tiny temperature variations in the cosmic microwave background, now cooled to a frigid 2.73 degrees above absolute zero. Famed cosmologist Stephen Hawking pronounced the finding the "discovery of the century, if not of all time."
Edwin Hubble set the stage for today's studies of galaxy formation when he discovered that the Milky Way was not alone. In the predawn hours of October 6,1923, at the Mount Wilson Observatory in California, he photographed a fuzzy, spiral-shaped clump of stars known as M3 1, or Andromeda, which most astronomers assumed was part of the Milky Way. He soon realized that within the clump he had found a tiny jewel: a star known as a Cepheid variable. This type of star has a wonderful property: Its brightness waxes and wanes like clockwork, and the longer it takes to vary, the greater the star's intrinsic brightness. That means the star can be used to measure cosmic distances. By comparing the true brightness of the Cepheid in M31 with its brightness as it appears in the sky, Hubble was able to determine the distance between Earth and the star.
He discovered that the star and the cloud, or nebula, in which it resided were a million light years away—three times the estimated diameter of the entire universe! Clearly this clump of stars resided far beyond the confines of the Milky Way. But if Andromeda was a separate galaxy, then maybe many of the other nebulae in the sky were galaxies as well. The known universe suddenly ballooned in size.
Hubble soon recognized that galaxies come in three varieties. Ellipticals, which converted most of their gas into stars long ago, resemble distorted footballs. Spiral galaxies, including our own Milky Way, account for two-thirds of the known galaxies in the universe. These galaxies have central bulges of old stars, just like an elliptical, but their cores are surrounded by disks containing slender, spiral arms still aglow with newborn stars. Our nearest spiral neighbor, Andromeda, resembles a Frisbee with a fried egg at its center. Finally, irregular galaxies are the plodders, apparently making stars at the same slow rate ever since they were born.
This diversity of galaxies is rooted in violence, according to Julio Navarro of the University of Victoria in British Columbia. Like Abel, Navarro relies on computer simulations to study galaxy evolution, but his work focuses on galaxies later in their life cycles, when they are prone to smash into each other and are chock-full of stars. Recent studies by Navarro and Matthias Steinmetz of the Astrophysical Institute Potsdam in Germany depict how collisions could have altered the appearance of a single galaxy as it made its way through some 12 billion years of cosmic history.
The first galaxy was a disk, Navarro believes, a consequence of the object's rapid rotation and the pull of gravity. As this disk repeatedly ran into and fused with other baby galaxies, the orbits of its stars became scrambled. The battered disk puffed into a swirling, sparkly ball of gas and stars—an elliptical galaxy. Then, as the galaxy slowly dragged in streamers of gas, the ball became the aging centerpiece of a bigger disk with spiral arms. Another collision erased that structure and created a larger ball. With each collision the galaxy altered its shape, like a lump of clay constantly being resculpted, but also growing bigger. The most popular version of the dark matter theory says that galaxies began small and grew over time through collisions and slow accumulation of material from their surroundings.
And these collisions aren't just things of the past, Navarro notes. Witness the Antennae galaxies, two galaxies caught in a cosmic tussle 63 million light-years from Earth. Their mutual gravity has pulled out two long streamers of luminous matter that resemble the antennae of a cockroach. Closer to home, the Andromeda galaxy, now hurtling toward us at 300,000 miles an hour (482,803.2 kilometers an hour), will merge with the Milky Way in several billion years, theorists predict.
It wasn't the orderly shapes of mature galaxies but the messy shapes of baby galaxies that captured the imagination of astronomer Chuck Steidel at the California Institute of Technology. His work has led to the discovery of more than 2,000 early galaxies—sometimes at a rate of a hundred a night—providing important data for theorists like Abel and Navarro. And it all began with a trek to a remote mountaintop in Hawaii.
As Steidel and three of his closest colleagues drove slowly up the narrow, bumpy road to the 13,796-foot (4,205-meter) summit of Mauna Kea, they knew this was their chance to crack open the secrets of the early universe. If the skies remained clear, they were about to observe the heavens with the largest visible-light telescope in the world, the Keck.
It was September 30, 1995, and Steidel, at only 32, hoped to accomplish what no one had ever done—detect in wholesale numbers galaxies so distant that the light they emitted more than 12 billion years ago was only now reaching Earth. That meant the galaxies would appear as they did when they were infants. If Steidel and his collaborators could find enough of them, these youngsters might reveal not only how galaxies first formed but also how they changed over time, and how they were distributed across the universe.
Until then astronomers hunting distant galaxies hadn't made much progress. They had found a few oddball objects that glowed extremely brightly, but they had failed to find the run-of-the-mill, remote galaxies thought to be prevalent in the cosmos. Most astronomers figured they would need bigger telescopes to find these faint objects. But Steidel had another idea: Maybe galaxies that hailed from the early universe had already been detected but no one had been able to pick them out from the thousands of other objects on sky maps.
Like a few other astronomers before him, Steidel realized that distant galaxies have their own signposts. They contain an abundance of hydrogen gas, as does the vast expanse of intergalactic space between them and Earth. When the ultraviolet light emitted by stars in galaxies is above a certain energy level, hydrogen gas absorbs it. The light never reaches Earth. So before Steidel and his collaborators ever dreamed of coming to Keck, they recorded galaxies that showed up brightly in red and green filters but were absent when viewed through an ultraviolet filter. They called these galaxies Lyman break galaxies, after Theodore Lyman, a physicist who pioneered studies of ultraviolet light in the early 20th century.
According to the color criterion, the faint galaxies Steidel's team had found prior to coming to Mauna Kea ought to be remote. But were they? To measure distance, the astronomers had to determine how much light from a galaxy had been stretched, or reddened, by the expansion of the universe. The greater this redshift, the greater the distance from Earth. A galaxy at a redshift of three, for instance, corresponds to a distance of about 12 billion light-years.
For faint galaxies, redshift can only be determined with a telescope as powerful as Keck. Now Steidel and his colleagues Mark Dickinson, Mauro Giavalisco, and graduate student Kurt Adelberger found themselves with two nights on the telescope. If they could demonstrate that their color method worked, they would have a foolproof way to find not just one or two distant galaxies but dozens—even hundreds.
Years before, Steidel and his collaborators had already picked out their first target. Residing in the constellation Eridanus, it was the brightest Lyman break galaxy the team had yet found. "We figured if we were going to be successful, it was going to be with this object," Steidel recalls. But he also knew that from Mauna Kea, the starlit body rose above the horizon for only an hour each night.
The fleeting hour that Keck stared at the galaxy, however, turned out to be enough. Just as Steidel had predicted, the spectrum revealed that the galaxy resided 12 billion light-years from Earth. Steidel was thrilled that his technique could find an ordinary galaxy so far away.
On the next night at Mauna Kea the astronomers attempted an even more ambitious feat. Taking full advantage of the power of the Keck spectrograph, they attempted to measure the distance simultaneously to several galaxies in the same patch of sky. To do so, they used a mask, a piece of aluminum about the size of a cookie sheet, which had several narrow slits carefully cut out. Each slit precisely aligned with the position of a target galaxy. With the mask in place, only the light from each target galaxy could enter Keck's spectrograph. By the end of that night the young astronomers had found 15 galaxies with redshifts greater than three.
On that night, slightly giddy from the high altitude, Steidel played for the first time at Keck the dreamy, lullaby-like music of the alternative rock band Mazzy Star. It would soon become a coda for each night Steidel observed at Keck and a special bond between him and Dickinson, whom he had met when they were both college disc jockeys at Princeton in 1980.
By 1997 Steidel's team had bagged another 250 Lyman break galaxies and an intriguing pattern emerged. To the surprise of the astronomers, those distant galaxies were strongly clustered in a way that revealed how dark matter is distributed. The first galaxies formed in the densest regions of the universe, which correspond on average with the densest regions of the cosmos today, where we find large galaxy groups and clusters. As time went on and gravity exerted its inexorable pull, regions of lower density also gave birth to galaxies, blazing with newborn stars.
Just as important was another discovery made by Steidel and Kurt Adelberger in 2001: Powerful winds were rushing out of the Lyman break galaxies, proving that there was more to the story of galaxy formation than dark matter. The winds, driven by supernova explosions, were so strong they enabled ordinary matter to temporarily escape the grasp of dark matter, which was unaffected by the winds. Not only did the winds clear out a vast bubble around their home galaxy, they carried hydrogen and other elements into surrounding space. The heavy elements, which could only have been forged inside stars, set the stage for future generations of galaxies.
"For a few weeks I dreamed about winds and thought about winds while I was eating my cereal in the morning and while I was in the shower and while I was Rollerblading to work," says Adelberger, now at Harvard. These winds added a layer of complexity to the story of how galaxies evolved from the simple universe of dark matter and elemental gases described by Tom Abel. Without such winds we can't easily explain the appearance of the visible universe today.
Beginning where Steidel's team left off, astronomer Sandra Faber of the University of California, Santa Cruz, is poised to break new ground in the study of galaxy formation. She and her collaborators hope to piece together how baby galaxies, like the ones found by Steidel, developed into the galaxies around us today.
Last March, wearing a navy blue jumpsuit that made her look more like an auto mechanic than a surveyor of the heavens, Faber strode through the chilly rooms of the Keck II observatory, which began operating in 1996 next to the first telescope. She had come to Mauna Kea to install the state-of-the-art Deep Imaging Multi-Object Spectrograph (DEIMOS) that she and her team had designed. The 20,000-pound (9,071.8-kilogram) device, which has to be slid in and out of position on metal tracks, can simultaneously analyze the light from as many as 130 distant galaxies.
"We're collecting the photo album of the life history of the universe for the first time," she said. "The baby pictures, the teenage pictures, the grown-up pictures." Astronomers are even taking snapshots of what the universe looked like before galaxies were born. If we used the birth of galaxies as our reference point, she said, then the hot and cold spots in the cosmic microwave background would be the prenatal pictures.
Faber is homing in on the process of galaxy formation from mid-childhood to early adulthood. At redshift three, galaxies were blobby and irregular. At redshift one, corresponding to a time when the universe was little more than half its current age, the shapes of galaxies cataloged by Edwin Hubble were beginning to fall into place. In between is a mystery interval from 12 to 8 billion years ago in which galaxies are notoriously hard to detect. During this largely uncharted interval galaxies matured, taking on their final mass and familiar shapes. A goal of DEIMOS is to open this interval to view.
"The spectrum of the night sky is the great enemy," she said, "an incredible picket fence of glowing emission lines"—the bright light emitted by atoms and molecules at sharply defined wavelengths. This picket fence in Earth's atmosphere overwhelms the faint infrared light from galaxies her team wants to study. There's one saving grace, however. The emission lines are narrow, while those from distant galaxies are much broader. With that in mind Faber's team designed DEIMOS to greatly expand, or disperse, the infrared spectrum. That enables the team to look between the pickets and focus on the light emitted by the galaxies.
That's when the fun begins. The brightness and shapes of the galaxies at different redshifts—and myriad other properties that can be observed thanks to DEIMOS—can indicate how the small, scruffy looking galaxies in the early universe formed the familiar galaxies that Hubble described in the 1920s.
Perhaps the most important of these properties is mass, Faber said. By measuring the mass of galaxies observed at different times in the universe, Faber hopes to trace the steps by which galaxies merge and grow larger. She would also like to learn why spiral galaxies, which are easily disturbed by collisions, are so abundant. The answer could be that in recent times spirals have grown by slowly drawing in material rather than through collisions. If her reasoning is correct, spiral galaxies should be forming stars at a gentle rate rather than in bursts that accompany collisions. Over the next few years, DEIMOS should provide the answer.
A few hours after Faber finished her work for the day, the domes of the twin Keck telescopes slid open and the instruments drank in the faint light from some of the most distant objects in the heavens. Down in Waimea, 48 miles (77.2 kilometers) away, two groups of astronomers were gathered inside an industrial-style low-rise building to relay instructions to operators on the mountain. Since 1996, a year after Steidel began his work, the telescopes have been directed from control rooms in this building.
In one room Arjun Dey of the National Optical Astronomy Observatory in Tucson, Daniel Stern of NASA's Jet Propulsion Laboratory, veteran observer Hy Spinrad of the University of California, Berkeley, and graduate student Steve Dawson were aiming the Keck I telescope at a catalog of faint galaxies, hoping to peer deeper than ever before into the universe—a billion years farther back in time than the galaxies found by Steidel. These are galaxies that glow brightest when they are observed through filters that allow only certain wavelengths of light to pass. The wavelengths correspond to a specific ultraviolet radiation emitted by hydrogen atoms that has been highly redshifted by the expansion of the universe. The filtered light was an indication, but not a confirmation, that the galaxies were located near the edge of the visible universe.
In the control room next door, meanwhile, Caltech astronomer George Djorgovski was also studying the distant universe. Using Keck II, he was trying to take the spectra of one of the most distant known quasars, the brilliant beacons that emanate from the cores of some galaxies. This quasar was so far away that to reach Earth, its light pierced regions so far back in time that they hadn't yet been blasted by radiation from the first generation of stars in the universe.
Back at the Keck I control room, Dey and his colleagues were staring at a bunch of squiggly black lines on the computer screen. After several hours of analysis, they came to a consensus. At a redshift of 5.74, the light that had fallen on the Keck telescope had left a galaxy known as LALA J142546.76 352036.3 more than 13 billion years ago. It appeared they had found the third most distant galaxy known. But after a final check, Dey and his collaborators smiled even more broadly and gave each other high fives. For on this night, March 13, 2002, the astronomers had found the second most distant galaxy known in the universe (after another galaxy discovered at Keck with a redshift of 6.56).
So what does it all mean? Have astronomers finally solved the riddle of how galaxies were born and evolved? Not quite, says William C. Keel of the University of Alabama, but astronomers are likely to put pieces of the puzzle together over the next decade. With mammoth new maps of the nearby cosmos, scientists today can study 13 billion years of galaxy evolution. But a veil still conceals what happened during the first, crucial period of galaxy formation, which astronomers have dubbed the Dark Ages. It began a few hundred thousand years after the big bang and ended perhaps a billion years later. During the first chunk of that time, the universe was truly dark. But later on, the first glimmers of starlight emerged and a telescope that has enough light-gathering power and is sensitive to just the right wavelengths, should be able to detect them.
A key task, already begun, will be to build a telescope to penetrate the veil. Keel and many astronomers are pinning their hopes on NASA's James Webb Space Telescope, the proposed successor to the Hubble Space Telescope, scheduled for launch about 2010. Equipped with a mirror capable of collecting six times as much light as Hubble, the telescope, with its advanced infrared and visible light instruments, will be able to detect objects much dimmer and farther away than those observed by any other telescope. That should give scientists the power for the first time to peer into the Dark Ages and to record the faint, warm light from some of the very first stars and galaxies, objects that can now only be seen in computer simulations like those on Tom Abel's laptop.
Until then the final frontier of galaxy formation awaits us, out in the darkness.