Giving Birth to Galaxies

Peering out to the edge of the universe, astronomers catch glimpses of galaxies in the making.

By Marcia Bartusiak|Saturday, February 01, 1997

Catching a galaxy in the making isn’t easy. It requires a great deal of cleverness and years of diligent searching. As a graduate student at the University of California at Berkeley in the mid-1980s, Mark Dickinson did not mind the hard work. He was one of the few astronomers lucky enough to be engaged in mapping the universe’s farthest frontier, the region billions of light-years away where galaxies appear to be frozen in their infancy in the distant past. It was a frustrating business, but also rewarding: understanding how galaxies were created is as important to astronomers as deciphering the origin of species is to biologists.

It could take hours or even several nights at the telescope to obtain just one good candidate, recalls Dickinson. That’s because he and his fellow observers were fighting the limits of their instrumentation. Given the time it takes light from a galaxy at the far end of the universe to reach Earth, the farther they peered into the depths of space, the farther they also saw into the past. The trouble was, they couldn’t see very much. Galaxies that resided more than a mere couple of billion light- years away were dim, fuzzy, and next to impossible to identify. There, wrote astronomer Edwin Hubble in 1936, we measure shadows and . . . search among ghostly errors of measurement for landmarks that are scarcely substantial. Astronomers have tried tracing the evolution of galaxies through those shadowy eons ever since, with little success.

But no longer. The trickle of data that distant-galaxy hunters once collected has now turned into a veritable geyser. Thanks to several key technological breakthroughs--the opening of the giant Keck telescope on Hawaii’s Mauna Kea in 1992, the improved vision of the Hubble Space Telescope, and advances in telescopic detectors--hundreds of young galaxies have now been sighted, with more being found each day. The early cosmos is fast becoming quite familiar. Says Dickinson, who now studies the far universe at the Space Telescope Science Institute, Since I’m used to scrabbling at the edge, I’m tempted to move on to something more murky.

Theorists have so much data now that we’re finding it hard to catch up, says University of Washington astrophysicist Craig Hogan. And with the flood of data, old ideas about galaxy formation are toppling. Particularly imperiled is the notion that virtually all galaxies came into existence at the same moment in the distant past, emitting a collective burst of light like some grand fireworks display. Astronomers had once seized on this explanation because it was the simplest. It also seemed to fit the available data: in the 1950s and 1960s, all galaxies as far back as astronomers could see (which wasn’t very far) looked pretty much the same as our own. Now, however, astronomers believe that galaxies condensed out of the primordial ocean of hydrogen and helium not all at once but continuously and vigorously, like a toasty fire, over a period of billions of years--through nearly half of the universe’s existence. Of course, not everybody agrees about precisely what is going on at the edge of the visible universe. At issue is the exact manner in which the galaxies came to be. Did most of them achieve their full size and identity fairly early, evolving only slightly beyond what they were at birth? Or did they take a walk on the wild side, starting from smaller bits that merged and coalesced gradually, and even at times swapping identities as easily as Imelda Marcos changed shoes?

For theorists such as Hogan, who endured the lean years of scant data, the debate is invigorating. He believes astronomy is now experiencing its third truly great moment this century. The first began in the 1920s when Edwin Hubble, peering through the biggest telescope of the time, the 100-inch reflector atop Mount Wilson in California, recognized that the Milky Way was not alone. After announcing that fact to the world on New Year’s Day 1925, Hubble went on to photograph hundreds of distant galaxies, classifying them by shape. About two-thirds of those are spiral galaxies just like ours--a bright central bulge surrounded by a spiraling pinwheel of gas and stars. Many of the other galaxies are denser and egg-shaped, or elliptical, and are largely filled with old stars. Ellipticals, Hubble saw, also tend to huddle together in rich clusters of hundreds and thousands. A smaller portion of galaxies, known as irregulars, are simply loose aggregations of stars immersed in rich pools of gas that do not have any definite shape. Hubble also observed that all the galaxies are moving outward at a speed directly proportional to their distance--an observation that would eventually be explained by the expansion of the very fabric of space-time from the moment of the Big Bang some 9 to 16 billion years ago (the exact age of the universe is still in dispute).

The second revolution culminated in the 1960s, when astronomers probed the heavens at nonvisible parts of the electromagnetic spectrum. Radio telescopes, for example, helped astronomers locate quasars, intensely luminous objects billions of light-years distant, which provided the first clues that the early universe was very different indeed from our local, rather humdrum galactic neighborhood. A quasar is thought to be a young galaxy that includes at its center a supermassive black hole formed from the stuff of millions of stars. As additional stray material falls into the black hole, it emits a tremendous light, as bright as a trillion suns. By the time this light reaches us from across the universe, it is very dim--so dim, in fact, that astronomers gazing through optical telescopes had completely overlooked it. Fortunately, most quasars also emit strong electromagnetic waves in the radio part of the spectrum, which makes them stand out when seen through radio telescopes. Only by using these radio emissions as a beacon for pointing their optical telescopes did astronomers finally get a good look at quasars.

At the time, many observers held fast to a specific strategy when it came to studying the early universe. They figured there was a precise era when galaxies were first constructed, when all those islands of new stars turned on in relative unison. It was a time when pockets of gas were gravitationally condensing--that is, forming stars--at tremendous rates, as gas supplies were at their peak. Astronomers therefore were looking for signs of a sudden eruption of light in the distant cosmos. They were searching for primordial galaxies that were making hundreds of stars each year. (The Milky Way now manufactures only about two new stars a year.) For years, they probed the far cosmos but came up empty-handed. They could say only that distant galaxies and clusters looked a bit bluer, a sign perhaps of heightened star formation. Young and massive stars, flush with energy, tend to put out more blue light.

Others, like Dickinson and his colleagues, had a different game plan. They tracked down particularly active galaxies with loud radio voices that could be heard across the universe. Perhaps the booming siren put out by a radio galaxy, they reasoned, was a sign that the galaxy was newly forming. As with quasars, the intense radio beam is believed to be radiating from a spinning black-hole dynamo lurking at the galaxy’s center. With the radio signal serving as a guide for pointing an optical telescope to the right position, observers could take a long exposure of the galaxy, since it was too faint to be noticed otherwise. But we had to look at dozens of candidates to find one that was really distant, notes Dickinson. It was exciting, but we ended up with only these exotic galaxies. The very same property that drew our attention--the strong radio signal--also made them abnormal. Astronomers weren’t sure they could understand the infant universe by studying only its most unusual specimens.

At this point astronomers faced a stark dilemma. After years of searching, the only evidence of bright objects from the early universe they’d found were a quasar here, a radio galaxy there. Perhaps, in retrospect, the fireworks theory was in some part wishful thinking: the very bright explosions of light it predicted were the only thing the telescopes of the day were powerful enough to detect from the early universe.

The new instruments that began to come on-line in the early 1990s gave astronomers another option. For the first time, they were able to look for far more subtle evidence of an alternative scenario--that galaxies were born slowly over many billions of years rather than all at once. To put it another way, astronomers had the means to look for primordial galaxies under the assumption that these galaxies might look pretty much like those near our Milky Way. We began to ask, What would an average galaxy, the kind we see around us today, look like some 7 to 10 billion light-years distant? says Caltech astronomer Charles Steidel. Pretty darn faint, was the answer. But Steidel had an intriguing technique for picking such dim workaday objects out of an already jam-packed nighttime sky.

A newborn galaxy sits at the center of a rich sea of gas (many times richer than the Milky Way’s current supply). Even though the millions of new stars in its core emit lots of bright, bluish light, the surrounding gas just sucks up all those ultraviolet photons, notes Steidel. The most energetic ultraviolet rays never get out of the galaxy. So if you use a spectrograph to break the galaxy’s light into its rainbow of separate colors, the resulting spectrum displays a gaping hole--a drop-off--where the high-energy ultraviolet photons should be.

That simple effect gave Steidel and his collaborators a way to start strip-mining the sky for infant galaxies. He knew that as the galaxy’s light travels through the cosmos, its waves get stretched with the universe’s expansion. The more distant the galaxy, the more its light is shifted to longer and longer wavelengths. Blue light turns redder, while red light waves move into the realm of the infrared. And that means the ultraviolet gap shifts as well. The position of the drop-off in the galaxy’s spectrum roughly pegs its distance, explains Steidel. It’s the poor man’s way of looking for early galaxies.

The idea was not original with Steidel, but he and his colleagues were among the first to apply it successfully. In 1991, using a telescope in Chile, they found about 20 potential baby galaxies. Each contender had an ultraviolet gap in just about the right position. But that was only a crude screening. To be certain, they needed more precise spectra. That couldn’t be done without the Keck telescope, Steidel says. In 1995 he transferred from mit to Caltech just so he could be closer to the giant instrument, with its 400-inch segmented mirror that’s four times wider than Hubble’s once-mighty telescope on Mount Wilson. Within a month of his arrival, Steidel checked out his candidates. Nearly all turned out to be quite distant, residing at a time when the universe was a mere one-fourth its current size and about one-sixth its current age. It took previous galaxy hunters more than a decade to gather such a sample from that era. Since then Steidel has amassed a collection of about 150.

More important, Steidel’s galaxies are not exotic. They’re proletarian, run-of-the-mill galaxies, stresses Dickinson, who worked with Steidel on the Keck observing run. Both believe they are seeing the initial cores of elliptical galaxies, as well as the bulges of yet-to-be spirals. (A spiral is thought to acquire its thin disk later, as the surrounding gas cools and settles down around the bulge.) And the population of objects that Steidel is counting in that far sector of the universe just about matches the population of bright galaxies that exists today. This seems to suggest that the major galactic components were in place within a couple of billion years of the Big Bang. That would put the rate of star formation in the average primordial galaxy into the reasonable range of 5 to 100 new stars a year. That’s fast by today’s standards, Steidel points out, but it’s not tremendously high.

Some complementary observations support Steidel’s picture. For the past ten years, Arthur Wolfe of the University of California at San Diego has been searching for primordial clouds of gas. Wolfe uses quasars as his tool. As a quasar’s brilliant light shines through the cosmos, some of its rays are absorbed by intervening gas clouds. By studying the absorption patterns--in effect, the shadows of these clouds--he can examine the gas between us and the distant quasar, as if he were inspecting some cosmic core sample billions of light-years in length.

Wolfe found that as you go farther into the past, more and more of the universe is made up of gas rather than stars and galaxies. By the time you arrive at the same era that Steidel observed--four-fifths of the way back to the Big Bang--he sees most of the mass tied up in neutral hydrogen gas. That mass, he found, is equal to the mass of all the stars in today’s universe. And that convinced many people we were seeing the progenitors of galaxies. The Keck telescope is powerful enough to allow Wolfe and his colleagues to observe how these gas clouds move. Even at this early, gaseous stage they appear already to behave like disks the size of the Milky Way, only thicker. This suggests healthy, big spiral disks are in place, says Wolfe. Which means, he believes, that the biggest galaxies formed fairly quickly and then coasted into the modern era largely intact. Smaller objects, the galactic fluff as he calls them, coalesced later. Assembly runs from big to small.

Or is it the other way around? That’s what evidence coming from the Hubble Space Telescope suggests. For ten days in December 1995, the Hubble trained its eye on one spot of the sky while it took a series of 342 time-exposure photographs. These images were then combined and computer- enhanced to produce the most deeply penetrating astronomical image ever taken, the Hubble Deep Field. This picture takes us billions of years back into time, maybe 80 percent of the way to the Big Bang, says Robert Williams, director of the Space Telescope Science Institute. It’s beautiful, yet also profoundly significant. It’s an archeological dig that allows us to see some 2,000 galaxies in different stages of development. Over the past year, astronomers around the world have been feeding on its data like hungry piranha.

James Lowenthal and David Koo, among others at the University of California at Santa Cruz, went out to the Keck to examine the Deep Field in detail. They saw many of the same galaxies that Steidel has observed but found them to be far more numerous than those seen today. Where did they all go? Either some of these objects have grown so dim that we can’t see them today, suggests Lowenthal, or they’ve merged. To Koo’s eyes, the objects appear rather small and blobby, as though they are smaller gaseous building blocks rather than early galaxies in their own right. Perhaps a galaxy does not condense out of a huge gas cloud, fully formed from the start. Instead it might fit together from smaller structures, like some cosmic-size set of Legos.

Additional evidence for this view comes from another Hubble observation, conducted by astronomers from Arizona State University and the University of Alabama. This group had the space-borne mirror take a set of exposures of a small point of the sky in the Hercules constellation. They uncovered 18 small Lego-like objects, all about 11 billion light-years distant and packed within a region only 2 million light-years across, about the distance between the Milky Way and its nearest spiral companion, the Andromeda galaxy. They believe they have succeeded in catching sub- galactic clumps in the act of merging into one or more galaxies. I don’t think that these objects are peculiar, asserts Arizona State astronomer Rogier Windhorst. I suspect we’ll see them all over the sky. They have already found similar structures in another, randomly selected field.

But critics warn that the light being collected from that far era essentially shows the objects in ultraviolet alone, which could be misleading. The compact bodies might simply be pockets of vigorous star formation embedded within hidden full-size galaxies, like bright lights strung over dark Christmas trees. To solve the puzzle, astronomers are trying to observe the motion of these bright objects to determine whether they are truly separate or simply bright parts of larger galaxies.

Should these faraway objects truly be small, it would please theorists no end. They favor a galaxy-forming recipe that calls for tiny units blending into ever-larger assemblies, from dwarfs to giants. Simon White, director of the Max Planck Institute for Astrophysics in Germany, has been trying to perfect just such a recipe for nearly two decades. According to his computer models, the first thing to form are small disks of gas, each about 3,000 light-years wide (just about the size of the objects that Windhorst and company claim they are seeing). These, in turn, join to form a galaxy’s central component, its bulge. Some of these bulges might run through their remaining gas very rapidly, in which case, unless they absorb new gas from another source, they turn into dwarf ellipticals. If gas is plentiful, and other conditions are right (they’re not crowded up against other objects, for example), they might enwreathe themselves in disks of gas and turn into giant spirals, where new stars continue to form for a longer time--longer than the rapid evolution suggested by Steidel’s and Wolfe’s observations.

And once they form, do galaxies retain their identity? No, answers White. It’s my bet that galaxies are continually changing. A spiral galaxy, for instance, could later encounter another spiral and merge to form a bulbous, giant elliptical. Indeed, some nearby ellipticals have been found to contain the vestiges of spiraling disks within them. White envisions a universe of great dynamism; only now are galaxies winding down as they deplete their fuel supplies. We are entering an era of cosmic ennui, as less and less gas becomes available for a galaxy’s star-forming needs.

The mystery of galaxy formation would be solved at once if astronomers could follow a particular galaxy through time, which, alas, is impossible. We can use a telescope as a time machine, but we’re either here or there, notes White. We can’t wait around for any individual galaxy to evolve. By adding more snapshots to their cosmic album, however, they can forge a link between one epoch and the next. Astronomers are also observing the universe’s subsequent stage of development--its adolescent years, about 5 to 7 billion years ago. So far observations from this cosmic era show ellipticals that seem remarkably old and undisturbed, as though they have stayed much the same for eons. But there are also ragged, strange-looking spirals merging and interacting, and bright dwarf galaxies bursting with star formation, as if they’re currently under construction. There is ammunition here for both sides of the galaxy- formation debate.

I know it sounds as if I’m waffling, says Koo, but I think both views of galaxy formation are right. Like the proverbial blind men and the elephant, Koo believes that astronomers have been feeling varied parts of the cosmos. Different astronomical techniques, he points out, will tend to pick out different types of objects. If you insist on taking the view that galaxies evolved in only one way, then you end up with a dilemma. It reflects our need to simplify, but I believe the universe can accommodate a richness of diversity, he notes.

Ideas may change as swiftly as new data are gathered. Another Hubble Deep Field survey is planned for the southern sky so that astronomers can use an array of giant telescopes coming on-line in Chile, the Southern Hemisphere’s astronomical mecca, to do follow-up work. And going deeper into space means peering at longer wavelengths, as the cosmic expansion stretches a galaxy’s light beyond the visible spectrum. Advanced infrared telescopes promise to push our cosmic vision even farther into the past.

Already there are hints of new discoveries. Both French and Japanese radio astronomers have detected carbon monoxide molecules in the distant universe, which could possibly be the residue of star bursts in an even earlier era yet to be surveyed. Like pole-vaulters, says Dickinson, we keep lifting that bar ever higher.

Comment on this article