It's past midnight in the dim telescope control room, but Dominique Naef's day has suddenly brightened. He twitches his computer cursor over a wavy line. "I like it," the Swiss astronomer says, beaming. "I like it a lot. Wow."
Fifty light-years away in the night sky, a star like our sun is doing a stately dance, stepping toward Earth and away again. From the La Silla Observatory in the mountains of Chile, Naef and his colleagues have stolen glimpses of the dance for months. But for much of that time their view was blocked by clouds, a foot of snow, and, this August night—midwinter in Chile—humidity so high that the telescope dome had to be shut to keep out frost. Earlier in the evening, between cups of espresso and cigarette breaks, Naef gloomily eyed a display of weather data. He feared another lost night.
Then the humidity dropped, and the telescope operator gave the go-ahead. Naef and Christoph Mordasini, a graduate student from Bern, huddled at their screens. They captured one more reading of the star's motion before, minutes later, the humidity shot up again and the operator called a halt for the night.
It's just another glimpse, but it's enough to turn a suspicion into a near certainty. The excited jiggle of Naef's cursor shows that the reading has fallen just where it should if an unseen planet is tugging the star to and fro. The next day the team leader, veteran planet hunter Michel Mayor of the University of Geneva, decides that it's time to announce the discovery. If it stands up to the scrutiny of other scientists, this planet, around a star called Mu Arae, will be a milestone in the quest for another Earth.
When Mayor and another colleague, Didier Queloz, found the first planet around another sunlike star a decade ago, it was a stunning feat. By now, astronomers tracking the wobbles of nearby stars have detected more than 130 alien planets. It's a strange harvest: gas-shrouded giants, mostly hundreds of times more massive than Earth, some in weirdly elongated orbits and others so close to their star that they circle it in days or even hours. But Mu Arae's planet, and two others reported at about the same time by U.S. groups, are far smaller than their predecessors and could be made largely of rock. With their discovery, the planet hunt has taken a turn toward the familiar.
These new planets are still no place for life as we know it. The planet around Mu Arae weighs at least 14 times as much as Earth—"an Earth on steroids," says one astronomer. It is scorchingly close to its star, completing an orbit every 9.5 days. But astronomers are convinced they will soon be finding solar systems where small, temperate planets like Earth could form and where some kind of life might flourish. "We're really on the doorstep of seeing systems like ours," says Debra Fischer of San Francisco State University, a member of the U.S. team that has found more than half the planets.
What they'll pick up first, they believe, are hints of giant planets in circular orbits far from their sun, like Jupiter—bodies that astronomers believe would raise the odds of Earthlike planets forming and surviving closer to the same star. The next step is actually taking a picture of an alien planet. At labs and mountaintop domes, engineers are at work on technologies capable of recording a planet's meager glow next to the glare of its sun. From the ground this optical wizardry could see a Jupiter-size planet. In space, aboard a multibillion-dollar mission called the Terrestrial Planet Finder that's scheduled to fly in a decade or so, it could pick up the light of a planet no bigger than Earth.
It's hard to overstate the excitement scientists feel at the prospect of seeing that faint blue dot. If it told of a watery, temperate place, humanity would face a 21st-century version of Copernicus's realization nearly 500 years ago that the Earth is not the center of the solar system. The discovery would show "that we're not in a special place, that we might be part of a continuum of life in the cosmos, and that life might be very common," says Michael Meyer, an astronomer at the University of Arizona. "To find oxygen, ozone, to see variations [in brightness] due to continents—that would be really exciting," says Sara Seager of the Carnegie Institution of Washington, who is developing techniques for interpreting that first glimpse of an Earthlike planet. "That's why I work so hard every day. Because that's what I want. I want to find that."
None of this would be happening if not for the winter nights that Didier Queloz, then a Ph.D. student at the University of Geneva, spent a decade ago at a telescope high on a plateau in southern France. Queloz and his mentor, Michel Mayor, were searching for hints of companions to nearby stars. They hoped to find dim, failed stars called brown dwarfs. But they were also trying their luck in a game that had disappointed many other astronomers before them: hunting for alien planets.
Their strategy was to break the light of each star into all its colors, producing a spectrum, banded with dark lines like a bar code. Each dark line indicates a wavelength of light soaked up by gases at the star's surface. If the star is moving toward or away from Earth, the lines shift to slightly shorter or longer wavelengths. It's the same effect you hear in an ambulance siren: As it approaches, the sound waves are squeezed to shorter wavelengths, raising the siren's pitch; after it races by, the sound is stretched out and its pitch drops. A rhythmic wavelength shift in a star's spectrum, back and forth, can mean that the star is heading toward us, then away. Something massive must be orbiting it, tugging it this way and that. The extent of the shift is a rough gauge of the object's mass, and the timing tells how long it takes to complete an orbit.
"The principle is simple," says Mayor, "but the devil is in the details." The telltale shift in the lines can be minuscule—no more than the width of a dozen or so atoms on a detector, for planets as small as his group's latest prize.
Back in 1994 astronomers also believed that it would take years, maybe decades, to see a star moving under the spell of a planet. The only planets able to yank their star fast enough to register on the instruments of the day would be the mass of Jupiter or more—hundreds of times heavier than Earth. And theorists believed that, like Jupiter, giant planets would only be found far from their stars in orbits taking years to complete.
To Queloz's astonishment, one of the stars he watched that winter was rocking back and forth every 4.2 days. It could have been a fluke, but just maybe it was a planet half the mass of Jupiter in an orbit no one thought possible—so close that the planet was practically skimming its parent star. With Mayor's encouragement, Queloz stuck with it, building up data until he was sure. "Maybe it was intuition, maybe it was luck," he says. "Maybe it was because I was young and naive."
Some astronomers thought the planet, called 51 Peg b for its parent star in the constellation Pegasus, was too bizarre to be real. But it turned out to be the first of an avalanche. After Mayor and Queloz announced it in the fall of 1995, other planet hunters who had come up empty took another look at their data, alert for fastpaced wobbles. They also expanded their search to hundreds of other nearby stars. More giants quickly turned up, some in searingly close orbits—"roasters," as some astronomers now call them—and others careering near and far with each orbit, on wildly eccentric, or oblong, paths.
These discoveries told of planetary turmoil—giants swept from remote birthplaces into hot, close orbits, and planetary siblings playing gravitational tug-of-war until one was flung into deep space. As Greg Laughlin, an astronomer at the University of California, Santa Cruz, says, "We're seeing the staggered, surviving remnants of systems that went berserk early in their history."
A few roaster planets have even been glimpsed more directly when they transit their star, crossing its face and dimming its light like a beetle crawling across a lamp. In the early years, skeptics argued that something other than planets might explain why the stars appeared to wobble. "The first transiting planet killed that idea," says David Charbonneau of the Harvard-Smithsonian Center for Astrophysics, one of the planet's discoverers.
Found in 1999 orbiting a star named HD 209458, that first transiting planet also gave astronomers their first reading of an alien planet's dimensions. Although lighter than Jupiter, this one is bloated to a diameter 35 percent greater. Subtle color changes as starlight shines through the planet's heat-swollen atmosphere are also yielding hints about what this alien world is made of—hydrogen, helium, and sodium, for starters—as well as signs that it is slowly evaporating in the heat. All in all, says Laughlin, "it's a bonanza if you can find a planet transiting a bright star." So astronomers are eager to find more. A group including Charbonneau has set up a network of small telescopes that scan thousands of stars night after night, watching for any transits—and recently detected another giant planet crossing the face of its star.
Roasters and eccentric giants are nothing like the planets we know, and the gravitational tug of these rampaging giants could threaten the survival of Earthlike planets. Yet planetary scientist Jonathan Lunine, flashing computer models of planet formation on a monitor in his University of Arizona office, isn't dismayed by these freakish worlds. "They're common in the best possible way," he says. "They're common enough that it looks like planet formation is a normal process." But they're not so common that solar systems like our own, where Earths can coexist with giants that move in wide, circular orbits like Jupiter's, won't turn up. So far, astronomers have found planets around only about 10 percent of the sunlike stars they've inspected. That leaves plenty of room to find alien Jupiters as searches become more sensitive.
Computer models by Lunine's colleagues show why that matters: Jupiters help Earths take shape. In the disk of gas and dust that surrounds a newborn sun, giant planets are thought to form first, in a million years or so. The leftovers provide the raw material for smaller planets. Dust clumps together into gravel, gravel to rocks, and rocks to hundreds of planetary embryos about the size of Earth's moon. But then the process grinds to a stop, at least in most computer models, because the embryos stay in tidy circular orbits like freeway drivers keeping to their lanes. They don't smash into each other, so they don't grow.
"That's where Jupiter comes in," says Lunine. With its powerful gravity, "Jjupiter is going to gradually pull on them and make their orbits more eccentric." he clicks through a few screen images, leapfrogging through tens of millions of years. Frame by frame a Jupiter's influence churns an orderly set of embryos into an unruly, colliding swarm. A handful of Earth- and Mars-size planets take shape in the turmoil.
Jupiter bestows a second blessing, the computer models show. At an Earthlike distance from a young star, the disk is too hot for water to survive in the planet-forming material, leaving any planet embryos bone-dry. The newborn planetary system's water would be locked up in icy embryos several times farther out—too cold for life. But a Jupiter would cause distant embryos to veer inward toward the star, delivering a generous splash of water to any newborn planets they collide with. On Lunine's computer the planet embryos at an Earthlike distance start out red, for dry. As Jupiter stirs the pot and the embryos give way to full-fledged planets, the color shirts to blue or green—sopping wet.
Later on, Lunine and others believe, Jupiters would act as bodyguards for these small, watery worlds. In a newborn planetary system, chunks of leftover rock and ice big enough to devastate an Earth would probably be on the loose for hundreds of millions of years. In our solar system,—Jupiter helped clean up the neighborhood. With its powerful gravity, it took hits from some of the troublemakers, flung some into deep space or the sun, and herded most others into the asteroid belt. The result is a protected zone where Earth survives.
So the road to another Earth, it seems, leads through another Jupiter. But detecting the wobble that would reveal a Jupiter means monitoring a star for a decade or more, while the giant completes its stately orbit.
By now, the wobble-watchers have been at it for long enough to be getting close. Last year a group led by Paul Butler of the Carnegie Institution of Washington reported a giant with a nearly circular, six-year orbit, about half the size of Jupiter's. Butler and his colleagues, including Geoff Marcy of the University of California, Berkeley, and Debra Fischer, may be just a few measurements away from announcing the real thing. "They probably have several very beautiful true Jupiter analogues in their data, and they're just waiting to get more data so they're really sure," says Laughlin.
There ought to be a more direct way to find these Jupiters: Take a picture of one through an extraordinarily powerful telescope. But that means finding some way to vanquish the glare of the star.
In a workshop under the bleachers of the University of Arizona's football stadium in Tucson, a telescope mirror the size of a small skating rink rests in a cradle. Frosty white and sightless, it's in the midst of three months of grinding, as a diamond-coated wheel carves the 27.5-foot (8.4-meter) expanse into a near-perfect shape for final polishing. When it's done, it will be trucked to the summit of Mount Graham, 70 miles (112.7 kilometers) to the northeast, joining an already completed twin. There the Large Binocular Telescope (LBT) is taking form. Starting in 2006, its dual mirrors will peer into space side-by-side like the saucer eyes of an owl, looking for Jupiters.
Sheer size will allow the LBT to see fainter objects than all but a handful of other telescopes in the world. But the real trick is the twin mirrors, which open the way to a feat of optical alchemy that can transmute starlight to total darkness.
That feat, called nulling interferometry, relies on the fact that light waves have crests and troughs, just like waves on water. By precisely aligning the light waves gathered by the two mirrors from a particular point in the sky, astronomers can overlap the wave crests from one mirror with the troughs from the other so that the light simply cancels out. Noise-canceling headphones use the same principle to deaden sound waves; with light, the result is a patch of darkness.
For this scheme to succeed, the light beams have to be guided and merged with exquisite precision. The payoff: blotting out the light of a star so that a giant planet, hundreds of thousands of times fainter and only a hairbreadth away on the sky, can be seen.
Even then the LBT probably couldn't see an exact counterpart of our Jupiter, dimly lighted by its distant sun. To be visible to an earthbound telescope, an alien Jupiter would have to be several times bigger or much younger—say half a billion years old instead of nearly five. Still warm from its violent birth, a young Jupiter would glow in infrared light, like a distant heat lamp.
Nulling has plenty of rivals as a Jupiter-finding tool. "This is a technical arms race," says Laird Close of the University of Arizona, who is pursuing a different approach. Working with telescopes in Arizona and Chile, he is exploiting subtle color differences between a star and a young Jupiter. With the right filters, he hopes to dim the star and make any planet pop out. Ben R. Oppenheimer of the American Museum of Natural History is trying something else. He has fitted a telescope on Maui with a set of precisely sized masks that physically block the starlight. "It's all just about deleting that star," he says.
None of this star subtraction would be possible without adaptive optics—a means of sharpening telescope images. In a perfect image a star would appear as a crisp point that could be neatly deleted, or just ignored, by astronomers searching for planets next to it. Instead, even the best earthbound telescopes ordinarily see a star as a fat smudge. Earth's atmosphere is the culprit: In the last few miles of light's journey from star to telescope, air turbulence scrambles and distorts it. Adaptive optics measures the scrambling with a special sensor, then sends the information to a flexible mirror that deforms and undulates many times a second—a frenetic funhouse mirror—to tidy up the image of the star.
Most of the world's big telescopes are already equipped with adaptive optics, and some planet hunters are trying their luck on existing systems, without any other tricks to mask out the star. A very young, red-hot giant planet just might be visible if it orbited at least ten times farther out than Jupiter. Even that "is sort of a long shot," admits Bruce Macintosh, an adaptive optics expert at Lawrence Livermore National Laboratory, who is carrying out his search at the giant Keck II telescope on Hawaii.
A group at the Very Large Telescope in Chile has already glimpsed what may be a newborn giant planet near a dim brown dwarf. But it would be harder to spot such a planet in the glare of a normal star. Macintosh says that after examining a hundred stars, "we have candidates"—faint spots that might be a planet but could also be a star in the background—"but nothing I would describe as an obvious smoking gun."
Macintosh and others are working on a new version of adaptive optics that could deliver smoking guns by the dozen. Called extreme adaptive optics, it would replace the hundreds of tiny pistons that reshape current flexible mirrors with thousands of smaller ones, and correct the light not hundreds but thousands of times a second. "That would get us a bunch of real planets," Macintosh says—hot young Jupiters at a Jupiter's distance. Each one would mark its star as a place to look for Earthlike planets once the search begins in earnest.
Some astronomers aren't waiting until planets as small as Earth are in reach. They're hoping that current tools will reveal worlds that might be habitable, although quite unlike our home. That's one mission of HARPS, the instrument responsible for the Swiss team's latest discovery. Its lair is inside the concrete base of a telescope dome at La Silla.
Outside is the daytime brilliance of northern Chile's mountainous desert. Inside, darkened corridors lead to a door that Michel Mayor opens with a magnetic card. Behind it is another door, massively padlocked. Mayor, proud father of HARPS, has no key. At the moment, no one on the mountain does. "People will say we are completely crazy—that we believe it is a Swiss bank," he jokes. In fact, the way is barred because it would be all too easy to perturb the exquisite temperature control, high vacuum, and optical stability that allow HARPS to sift starlight for hints of planets smaller than any yet detected.
Silvery optical fibers snake into the innermost room, carrying starlight from the 3.6-meter (11.8-foot) telescope above. HARPS, sealed in a ten-foot-long (3.1-meter-long) vacuum tank, splits the light into a spectrum and monitors thousands of lines for wobbles. It's the same strategy that has yielded nearly all the planets found to date. But HARPS, up and running since late 2003, brings many times more precision to the task than Mayor's earlier instruments—precision enough to tell when a star tens of light-years away is moving toward or away from Earth at a speed no greater than a walk.
That sensitivity was key to glimpsing the Mu Arae planet, so small it exerts only a feeble tug on its star. It just might allow HARPS to detect similar planets—say 20 times Earth's mass, or roughly the mass of Neptune—in orbits the size of Earth's. No one knows what such worlds would be like, but planets like Mu Arae's have raised hopes that they exist. The Neptune we know is a ball of ice and rock near the edge of the solar system, but a similar planet closer to its star might resemble an oversize version of Earth, with a rock surface. Or it might have an ocean hundreds of miles deep. "We can dream," says Queloz.
So can Mayor and Queloz's competitors. Weeks before the Swiss team was sure of the Mu Arae discovery, two U.S. groups had quietly firmed up the case for other small worlds. Barbara McArthur of the University of Texas's McDonald Observatory found a planet weighing as little as 14 Earth masses—as small as the Mu Arae find—racing around the star 55 Cancri every 2.8 days. Paul Butler and his colleagues added their own bantamweight, at 21 Earths. Now, at the Lick Observatory near San Jose, California, their group is building a special-purpose telescope aimed at finding Neptune-size worlds far enough from their star to be habitable.
Robotically controlled for efficiency, the 2.4-meter (7.9-foot) Automated Planet Finder will capture every glimmer of starlight with mirrors plated with silver instead of the usual aluminum. Next year Debra Fischer will set it to work inspecting a hundred stars night after night for hints of worlds compact enough that they just might host life on a solid surface or in a deep ocean.
A planet as small as our own, however, will remain out of reach for both teams. That's because stars pulse and roil, creating surface motions that would make it impossible to detect a star's tiny drift—barely a crawl—under the spell of an Earth. But there are other ways to parse starlight for hints of real Earths.
Forty years ago William Borucki helped design heat shields for the Apollo moon missions. Not far from his office at NASA's Ames Research Center near San Jose, parachutes for the Mars landings in January were tested in the silvery wind tunnels that run between buildings like oversize air ducts. But Borucki's ambitions have vaulted far beyond the solar system. At an age when many people think about retirement, he's planning a four-year, 400-million-dollar space mission to hunt for Earth-size planets. "This is really pure exploration," he says. "This is sending the Nina and Pinta out to see how many dragons there are on your way to India." Borucki's vessel is Kepler, a space telescope half the size of the Hubble but designed for the single purpose of planet finding. To be launched in 2007, it won't capture light from other Earths. Instead, from a vantage far beyond the moon, it will chase their shadows. Like the astronomers watching from the ground for giant planets crossing the face of their stars, Kepler will watch 100,000 stars in a broad patch of sky—as wide as two hands held at arm's length, says Borucki—for a dimming that signals a planet the size of Earth, or even smaller. "I'm looking for this dip that repeats, and repeats exactly," he says.
The dip due to an Earth would be so tiny—less than one part in 10,000—that it could be seen only from space. Even then more than 99 of every 100 Earths would elude Kepler, because it will be able to detect only the half a percent that have orbits aligned so they pass directly in front of their stars. Spacecraft vibration and "sunspots" on the target stars could throw off the brightness measurements. But Borucki is confident in his instrument. The real uncertainty, he says, is "whether there are no Earths at all or tens of thousands" around the stars Kepler will watch.
He's quick to add that everything astronomers have learned about giant planets so far suggests these smaller siblings are out there too. By 2011 or so he expects Kepler to have detected a few dozen Earth- or Mars-size planets set just far enough from their stars to be comfortable for life. The shadows won't say much about what these worlds are like. But the flicker of distant Earths would galvanize the next step in the quest: actually capturing their light.
"That's a hard, challenging technology problem," says Charles Beichman, the project scientist for the Terrestrial Planet Finder (TPF) mission. He's putting it mildly. Seeing alien Jupiters is proving hard enough, and a planet as small as Earth would be far more elusive, snuggled right next to a star shining ten million to ten billion times brighter, with an unknown amount of interplanetary dust adding its own distracting glow. Earth hunters like Beichman compare the challenge to that of seeing a firefly hovering next to a lighthouse searchlight 3,000 miles (4,828 kilometers) away—with a little fog rolling in.
The key is to blot out the lighthouse while leaving a clear view of the firefly. The project can draw on the technology already being tested for spotting Jupiters from the ground. But the stakes are vastly higher, because TPF will cost so much—upwards of a billion dollars—and generate so much anticipation. "You'd like not to have zero [Earths]," says Beichman, adding wryly, "Zero is a bad number, because you get called up in front of Congress and asked how come you spent a billion dollars and didn't find anything."
For now his team is betting on a pair of technologies. First to fly, in 2014, would be a single telescope, with a mirror perhaps 21 feet (6.4 meters) across and clever masks to banish starlight from the spot where a planet might appear. A nulling interferometer—multiple smaller infrared telescopes that merge light to create an optical dead zone blocking the star—would follow by 2020.
A possible joint project of NASA and the European Space Agency, the interferometer telescopes would fly independently, in a small fleet trading light beams. Free-flying telescopes could fan out across hundreds of feet, sharpening their combined view. But keeping them in near-perfect formation so that their beams mesh precisely would add to the technology challenges.
How masks, nulling, and other tools perform in the ground-based Jupiter hunt will help the TPF team refine their plans. Kepler's discoveries will feed into the planning as well: Knowing how common Earth-size planets are will help scientists decide how many stars to inspect.
By the time TPF is launched, astronomers should also have an idea which stars are the best prospects. As early as 2009 NASA hopes to launch another ambitious mission, called the Space Interferometry Mission, or SIM. SIM, like one version of TPF, is an interferometer—a set of small telescopes, in this case mounted on a single spacecraft. The 900-million-dollar mission is not designed to see an Earth directly but to monitor the positions of thousands of stars with painstaking precision. "Picture an astronaut on the moon holding a nickel edge on," says Stephen Edberg, a SIM scientist at the Jet Propulsion Laboratory (JPL) in Pasadena. SIM should be able to pick up a change in a star's position on the sky no greater than that nickel's thickness.
Suppose SIM sees a nearby star sidling back and forth by that tiny amount over many months. Like the oscillations that planet hunters watch from the ground, this subtle shimmy would imply an unseen dance partner—a planet that might be only a little bigger than our own. That star would be a prime target for TPF. "If SIM can detect a planet," says JPL's Michael Shao, the project scientist, "it can tell TPF when and where to look" for the light of an Earth.
If TPF does see that faint point of light, scientists will wring every bit of information they can from it. They'll want to learn whether that distant world has an atmosphere and a surface anything like Earth's̵which means knowing what our own planet would look like if its light were reduced to a single point.
One clue, says Nick Woolf of the University of Arizona, comes from the pale glow seen on the dark part of a crescent moon. Called Earthshine, it's sunlight reflected from Earth onto the moon's rough surface and then back again. Earthshine is a jumble of light from our atmosphere, clouds, oceans, and continents, just as the faint light from an alien Earth would combine light from all of its surface features and atmosphere.
Woolf and others have found they can unscramble Earthshine to see the imprint of gases in our atmosphere, the color of the oceans, and the blue of the sky. They can even discern a signature of vegetation called the red edge: a jump in brightness at the boundary between red light—which plants absorb—and infrared, which they reflect.
The shine from another Earth could be millions of times fainter, and would be far less informative, at least to a first generation planet finder. But if that far-off world has an atmosphere like ours, TPF should be able to see signs such as carbon dioxide and water vapor. If the atmosphere is rich in oxygen or its chemical cousin, ozone, TPF should detect it. And that would be an epochal discovery. As astrobiologist Vikki Meadows of JPL explains, "That would give us a good clue that something funny is going on, because we don't think we can create large amounts of oxygen without life."
The gas's source might be just a green tinge in an otherworldly ocean, or a crust of microbes on alien soil. But that first inkling of life across the light-years would amount to a curl of smoke on the horizon, a first hint that the universe may not be as lonely as it has seemed. Nick Woolf is 72, a mentor to a new generation of astrobiologists. When he began working on concepts for an Earth finder nearly 20 years ago, worlds beyond our own were no more than a dream. "Now," he says, "finally there has come a chance to really see what's out there."