All Aglow in the Early Universe.

Was the infant cosmos a star-making machine?

As astronomers peer ever deeper into space and farther back in time, they hope to finally come upon the epoch when star formation had just begun to bloom. But if astronomer Kenneth M. Lanzetta is correct, researchers will have to look farther back than most had ever imagined. Stars and galaxies may not have assembled gradually, as the most popular theory of cosmology suggests,

Instead, galaxies may have burst upon the scene early on. The infant universe wasn't merely aglow with starlight, it was a blazing inferno, propose Lanzetta of the State University of New York at Stony Brook and his colleagues.

According to this scenario, just a few hundred million years after the Big Bang, the cosmos was already churning out stars at a rate perhaps 10 times faster than it does today. Since stars reside exclusively in galaxies, that suggests that the very first galaxies were brighter and more massive than previously estimated, he says. Galaxies may have formed wholesale as behemoths, rather than built themselves up over time.

Lanzetta admits that such a model of galaxy formation hasn't been popular for several decades, since it violates a widely held view of the universe in which most mass is made of an unseen, slow-moving type of matter called cold dark matter, in the simplest cold-dark-matter theories, the dark matter is the first to clump together under the influence of gravity, and the ordinary material that comprises the visible, starlit parks of galaxies takes much longer to assemble.

"Many of the people who believe that the cold-dark-matter model has been quite successful would find [Lanzetta's proposal] heretical," says Alan Dressler of the Observatories of the Carnegie Institution of Washington in Pasadena, Calif. "If things were really that monolithic and formed that early, why should we be seeing all this activity later in the universe?" he asks.

Although several predictions of cold-dark-matter models have proved correct, observations don't rule out alternative models of galaxy and star formation, Lanzetta asserts.

He presented his controversial findings last month in Baltimore at a Space Telescope Science Institute symposium celebrating the 10th anniversary of the Hubble Space Telescope.

In interpreting observations of the early universe, astronomers have neglected an important fact, Lanzetta maintains. Most of the light emitted by the very first galaxies in the cosmos is much too dim to be seen today. Objects that were bright long ago appear faint now, and less brilliant objects are entirely invisible. The true intensity of ultraviolet radiation, a measure of starbirth, from extremely distant and very young galaxies has never been accurately recorded, Lanzetta asserts.

Using a novel method to account for the unseen radiation, he finds that 10 to 20 percent of all the stars in the universe today might have been born in a huge, early burst. "There is as yet no direct observational evidence that rules out a very early epoch of intense star formation," he says.

Lanzetta says the critical issue is what can and can't be seen at great distances in the cosmos. Astronomers gauge distance by measuring a quantity called redshift--the amount by which cosmic expansion stretches, or shifts, the light emitted by an object to redder wavelengths. The more distant the object, the greater its redshift.

It turns out that when an observer peers at a distant, and thus highly redshifted, galaxy, the intensity of the light it emits appears drastically reduced. The expansion of the universe contributes to the dimming. For instance, an observer on Earth detects particles of light, or photons, from a distant galaxy at a far lower rate than that at which the galaxy emitted the photons. This well-known effect, time dilation, causes far-off galaxies to look even fainter than their distance would suggest.

As a result of this and other consequences of the universe's expansion, brightness decreases rapidly as a function of distance. Astronomers call this effect cosmological brightness dimming.

Because of it, a distant galaxy may be much bigger than it appears in a telescope image. Only the central, brightest parts of a distant galaxy show up. Cosmological dimming puts intrinsically faint, outer regions below the telescope's detection threshold. In contrast, a telescope can record both the intrinsically bright and faint regions of a nearby galaxy, Lanzetta notes.

As a result, an observer viewing a nearby galaxy can correctly infer its minimum rate of star formation by measuring its output of ultraviolet light. That same method, however, would seriously underestimate the ultraviolet light emitted by a distant galaxy--and in turn, its minimum star-formation rate.

"This is exactly what has gone wrong with all previous measurements," Lanzetta asserts. "Astronomers neglected the ... dimming effect and thought they were measuring all of the light, whereas in fact they were missing most of the light." He calculates that telescopes can't directly measure the total amount of ultraviolet light from galaxies that lie farther away than about 5 billion light-years, roughly halfway to the edge of the observable universe.

"I think there's certainly merit to the idea ... that we're missing some fraction of the light output from distant galaxies," says Wendy L. Freedman of the Carnegie Observatories. "Certainly, people are aware of cosmological dimming, but trying to ascertain the form of the star-formation history over time is very difficult," she adds.

Astronomers can't easily calculate the full effects of dimming since they don't know how galaxy properties change over cosmic time or what fraction of the total galactic population they're seeing when they look at the early universe, she notes.

In computer simulations, some researchers have tried to assess the effects of dimming by placing nearby galaxies at long distances and measuring the amount of light that goes undetected, notes Robert C. Kennicutt of the University of Arizona in Tucson.

"The problem is that the amount of light lost is strongly dependent on the structure of the distant galaxies, and of course, that is what we're trying to measure," he says. "The current evidence suggests that galaxies were more compact in the past and dimming is less of a problem, but this is not established with certainty."

Rebecca A. Bernstein of Carnegie, along with Freedman and Barry F. Madore of the California Institute of Technology in Pasadena, tried to determine just how much light astronomers are missing. Using the Hubble Space Telescope, the team imaged in visible light a patch of sky that provides a relatively clear window outside our galaxy. Simultaneously, they took spectra of part of the same region of the sky with a 2.5-meter telescope at Las Campanas Observatory in La Serena, Chile.

Because light scattered by dust in the solar system has a particular spectral fingerprint, the researchers could identify this radiation and subtract it from the Hubble images. What remains is light associated with specific, detectable galaxies combined with radiation from galaxies too faint to be individually detected.

They found that the radiation is about three times the amount that can be measured from detectable galaxies. Part of the extra radiation, Bernstein says, belongs to the galaxies that have already been observed but whose light is so diffuse that astronomers overlooked it. In addition, some of the radiation comes from galaxies that have been completely overlooked because they are too faint and diffuse to show up as individual sources on a telescope image.

So, what's an astronomer to do? The most conservative answer, Lanzetta says, is nothing--at least until telescopes become thousands of times more sensitive.

He and his colleagues are now suggesting, however, a way to indirectly measure the light that's too faint for current telescopes to record. The researchers rely on a well-documented relationship between the density of gas and the rate of star formation in nearby galaxies. Kennicutt showed several years ago that among nearby galaxies, those with the highest density of gas have the highest rate of starbirth.

Lanzetta assumes, but can't prove, that a similar relationship holds in distant galaxies. Astronomers can measure gas density relatively easily in a distant galaxy, that is, if the galaxy happens to lie along the line of sight from Earth to a more distant quasar. As the brilliant quasar light passes through the galaxy, gas absorbs certain wavelengths of light. The amount of absorption reveals the density of the gas.

In this way, says Lanzetta, he uses gas density as a stand-in for the missing radiation. Although his technique is speculative, he notes that among relatively nearby galaxies, in which astronomers can measure star formation both by recording ultraviolet light and by using gas density, the two methods seem to agree. The apparent match, Lanzetta says, "might indicate that [gas-density measurements] aren't an entirely unreasonable way to go."

Kennicutt notes, "The imponderable in all of this is knowing what the distribution of gas in galaxies was at these early epochs.

"Using the gas density ... to infer the missing star formation is a clever idea," he adds. It should work at high gas densities but will be less accurate at lower densities, where star formation declines sharply, he cautions.

Lanzetta and his colleagues have applied the new technique to some 2,700 galaxies, a mixture of nearby and distant objects. Of this group, 69 appear to have redshifts higher than 5, indicating they lie farther than 12.7 billion light-years from Earth. Of these, 17 may have redshifts higher than 10, corresponding to distances greater than 13.2 billion light-years.

Lanzetta cautions that because some of the galaxies are extremely faint, the measurements are highly uncertain. "Some of these galaxies could lie much closer," he notes.

Accounting for these uncertainties, Lanzetta and his collaborators estimate a rate of star formation that increases roughly tenfold as they probe back in time from today's universe to more than 13 billion years ago. Because the researchers don't take into account dust, which would hide starbirth of distant galaxies, the rate could be even higher, Lanzetta notes.

Although other scientists have recently suggested that galaxies were already producing stars at an enormous rate when the universe was only 4 billion years old (SN: 2/13/99, p. 103), they say they're skeptical of Lanzetta's claim that the rate was significantly higher at even earlier times.

If stars formed so early and plentifully, notes Joseph I. Silk of the University of California, Berkeley, they would have generated large quantities of elements heavier than helium. "I think this would create great difficulties for our understanding of chemical evolution," he says, since the oldest stars have extremely low abundances of heavy elements.

"If Ken [Lanzetta] has convincing evidence that a greater fraction of stars than 10 percent formed before redshift 5, then, yes, I would agree that cold-dark-matter models of galaxy formation are in big trouble," says Carlos S. Frenk of the University of Durham in England.

Both theorists and observers, however, have raised a host of concerns about Lanzetta's proposal.

"While cosmological dimming is surely affecting our view of the distant universe, quantifying such an effect is difficult. It's difficult to correct for something you don't actually see," cautions Piero Madau of the University of Cambridge in England.

"I agree with basically everything that Ken is willing to make a stand on," says Harry C. Ferguson of the Space Telescope Science Institute in Baltimore. While acknowledging that star-formation rates could have been high in the early universe and that astronomers may have overlooked much of the light from distant galaxies, he says that he's skeptical of many of Lanzetta's methods.

Lanzetta's work, Ferguson notes, assumes that all gas above a certain density forms stars at the same rate. This relationship might not hold true at very early times in the universe, he says.

In his own studies, Rodger I. Thompson of the University of Arizona has used a near-infrared camera on the Hubble Space Telescope to observe starbirth in extremely distant galaxies. Out to the most distant objects that he can detect, which he estimates to have redshifts of about 7, he deduces a relatively constant star-formation rate rather than one that was higher at earlier times. He says that taking dust fully into account may reveal a more level rate of starbirth over time.

Kennicutt notes that most stars may still have been produced after the first epochs of the universe. "The small amount of time elapsed between redshifts 2 and 10 [an interval of about 2.2 billion years] suggests the total mass of stars formed during that interval of time might still be relatively small," he says. "Of course, we have never [directly] observed galaxies with redshifts greater than 5.6 or so, so inferring the properties of gas distributions, star-formation rates, et cetera, for [more distant] objects is highly speculative."

For now, the debate seems wide open. Although Freedman doesn't believe that astronomers have a solid grasp of the star-formation rate over the history of the cosmos, "I suspect that a rate that rises steadily, as claimed by Ken, is probably an extreme," she says. "But I agree with his premise that we're not detecting many galaxies as a result of surface-brightness-dimming effects, and that the rate could well be higher than [that] inferred from galaxy counts to date."

Lanzetta's work "is very provocative," says Dressier. "It's too early to know if he's right or wrong."