Gliese 876 is a modest star, just one-third the mass of our sun and only 15 light-years away, but it has a history-making planetary system all its own. In 1998 a team led by Geoff Marcy of the University of California at Berkeley detected the first sign of something interesting there: a giant planet, twice the mass of Jupiter, circling Gliese 876 once every two months, its gravity yanking the star back and forth at the speed of a jet plane. Three years later the same group found a second planet, half the mass of Jupiter and closer in, pulling the star around at the speed of a race car. Although the planets are too faint to be seen directly, their motions cause the star’s spectrum to wobble back and forth across the digital detector of an astronomical telescope.
In the past decade, announcements of Jupiter-size planets have become commonplace; about 300 of them have been found so far. In 2005, however, with the help of improved detection software, Marcy’s team turned up something else orbiting Gliese 876—something truly new. This invisible object added one more regular component to the star’s motion, like the third note, faint and high, of a piano chord. It was another planet, orbiting in just two days and pulling on the star much more gently, not at jet plane or race car speeds but at a speed a man could run. This planet, dubbed Gliese 876 d, is clearly no Jupiter, Marcy realized. It is no more than seven or eight times as massive as our own: a “super-Earth.” Until then, all the known exoplanets (planets circling other stars) were big and gaseous, but this one is probably made of rocky materials—the first world like ours found in an alien solar system.
Gliese’s super-Earth lies so close to its star that it has just about no chance of being inhabited. If it has an atmosphere at all, it probably consists of dense steam, says Greg Laughlin of the University of California at Santa Cruz, a member of the discovery team. But if we can find one rocky, Earth-like planet right in our galactic backyard, surely there must be many more. Already, the Swiss astronomers who in 1995 discovered the first Jupiter-like exoplanet—and who are the great rivals of the California group in the exoplanet hunt—said in June that they had identified not one but three super-Earths orbiting a single star 40 light-years away. The smallest is just four times as massive as Earth. “We’ll find an Earth-mass planet by 2010,” Laughlin predicts, “and an Earth-mass planet that’s potentially habitable by 2012.”
And yet we still won’t have found a true second Earth. The hallmark of Earth, after all, is not its mass, nor its rockiness, nor the fact that it is potentially habitable. The hallmark is that it is actually inhabited. These days, nobody doubts that there are other reasonably cool, rocky planets out there among the 100 billion stars in the Milky Way. Everything astronomers have learned about how stars and planets form says there must be. But is there life on any of those rocks, and if so, can we detect it? “That’s not going to happen from Earth,” Laughlin says. “It has to happen from space.”
In space, above our atmosphere, stars do not twinkle; in space a telescope is also beyond day and night and can thus stare at the same star for weeks on end, gradually teasing from its light the barely perceptible but regular flickers caused by a small orbiting planet. A French satellite called Corot, the first space telescope devoted primarily to looking for rocky planets, is in orbit now. An even more capable American mission, Kepler, will be launched in April. It is expected to find hundreds of Earths, including the first ones orbiting stars like the sun at distances like that of our own Earth. Then, in 2013, NASA will launch a giant infrared telescope called the James Webb Space Telescope. An all-purpose observatory, the Webb was not designed to follow up on the discoveries of Corot and Kepler. But if pushed to the limit, it just might be able to provide the first indication of life—a telltale molecule, such as oxygen, in the planet’s atmosphere—on a super-Earth circling another star. By 2014 headlines could be announcing the first tentative evidence of life beyond our solar system.
One rainy Tuesday afternoon last November, Annie Baglin, the chief scientist of Corot, sat at the window of a café near the Place Denfert-Rochereau in Paris, drinking tea. It had not been an easy day. French railway workers, striking over their retirement benefits, had shut down the commuter trains, preventing Baglin from reaching her office on the suburban campus of the Paris Observatory. The railway workers and other civil servants were marching down the Boulevard Montparnasse, near Baglin’s home, brandishing bright red flares that filled the air with chlorine-scented smoke; off the boulevard, platoons of shield-toting, armor-wearing riot police stood nervously at the ready. Arriving late for her rendezvous at the café, wearing the dark pink coat she had said would make her recognizable, Baglin explained that her car had been towed—apparently the traffic wardens were not on strike. From the café Baglin would be proceeding to her dentist to have a tooth extracted. On the plus side, her spacecraft was performing beautifully.
As Baglin launched? into the story of the little spacecraft that could, in principle, find many rocky planets, her high, thin voice sometimes disappeared into the noise of the sirens outside. She is a shortish woman of 70, with close-cropped gray hair and a warm, no-nonsense demeanor—her parents were both schoolteachers. A brief profile of her on the Paris Observatory Web site is titled “Annie Baglin—Never Say Die.” Getting Corot to the launchpad, she explained, was a long, hard slog, marked by bureaucratic near-death experiences.
She never intended to be a planet hunter. In the mid-1980s she and her colleagues proposed a space telescope to do stellar seismology—to study the inner workings of stars from vibrations on their surface, much as seismologists study Earth’s interior by analyzing earthquakes. The French and European space agencies were noncommittal about the idea. Then came 1995 and the announcement of the discovery of the first exoplanet, by Michel Mayor and his colleagues at the Geneva Observatory. Baglin and everyone else immediately realized that a spacecraft designed to detect the light fluctuations caused by starquakes might also be able to detect a planet. Suddenly, Baglin says, “we were very much sought after. In hindsight, one can say that if it hadn’t been for the discovery of exoplanets, we never would have been approved to do Corot. That’s what sold it.”
Launched in December 2006, Corot is thus a 1,300-pound spacecraft that does two very different things. No telescope yet exists that can take a picture of even a giant exoplanet; astronomers compare the task to taking a picture of a firefly next to a searchlight thousands of miles away. Mayor and his colleagues showed instead that it was possible, through a technique called astrometry, to detect the slight wobble in a star’s light caused by the gravitational pull of an orbiting planet. Most of the 300-some exoplanets discovered since have been found that way. But Corot relies on a different technique that has lately come to the fore in ground-based searches as well. Called photometry, it detects the slight but regular dimming in a star’s light when a planet transits in front of it.
What the search for planetary transits has in common with the observation of starquakes is the need to stare at the same stars for a long time—long enough to detect very slow vibrations or to detect at least three transits of a planet. Otherwise, you can’t be sure it was really a starquake or a planet you saw, and not random fluctuations in the starlight. Corot stares at the same spot in the sky for 150 days before switching to another. “Corot is Zen,” Baglin says. “Once we’re set up, we don’t move. We don’t even breathe.”
The spacecraft’s 27-centimeter (10.6-inch) telescope monitors up to 12,000 sunlike stars at once. Getting a big sample is crucial because only one in a hundred of those stars that do have planets will be oriented so that the passage of the planet in front of the star is visible from Earth. The precision of the telescope’s measurements has exceeded its makers’ hopes. “If Corot were to observe the million lightbulbs that shine along the Champs-Elysées at Christmas,” said a press release from the Paris Observatory a few days before Christmas in 2007, “it would be able to detect whether a single bulb was flashing.” That parts-per-million sensitivity should allow Corot to detect the dips in a star’s light caused by a transiting planet with a radius just twice that of Earth—and perhaps an even smaller one, provided its orbit is tighter than Mercury’s, so that the planet completes three transits during the 150-day viewing period.
Not long after the launch, the Corot science team, including Baglin, published a description of the mission. It concluded with this prediction: “The first confirmed terrestrial planets are expected in the spring of 2008.” By last spring, however, the Corot team had announced only two new “hot Jupiters” and one unconfirmed super-Earth, with 40 more candidates in the pipeline. Seeing transits is not enough; periodic dips in a star’s light could be caused by a small companion star too dim to detect directly. To confirm a planet, Corot’s candidates have to be observed from the ground using the wobble technique, which determines the mass of the transiting object; a planet will be much lighter than a companion star. But competition for telescope time on the ground is fierce, especially with so many planet hunters around. “It’s a real bottleneck,” Laughlin says.
Baglin has little patience for impatience or for the pressure on her to announce discoveries quickly. “Finding these things isn’t like finding the nose on your face!” she exclaimed, shortly before leaving the café last November to head for the dentist. The Swiss astronomer Mayor gathered data for 20 years, she pointed out, before announcing his first exoplanet. “So when people tell me, ‘You haven’t got any results,’ when we’ve only been in orbit for a year, I say, ‘Stop! Have mercy!’ In three years we’ll have results indicating how common small planets are. Big planets we know we’re going to find. We’re looking for the little ones. Are they there, or aren’t they?”
What about figuring out if one is inhabited? “To me, that’s not the big question,” Baglin said. “Understanding the universe in its totality interests me more than looking for life on a planet like Earth around a star like the sun, which is the declared goal of our competitors. That’s not uninteresting, but it’s not what excites me. I find that very anthropomorphic.”
Corot’s competition is Kepler, and Baglin’s is an astrophysicist named William Borucki at NASA’s Ames Research Center in Mountain View, California. In 1984 Borucki published his first description of the transit technique they would both end up using. At that time Baglin was having her first glimmerings of what would become Corot but had no notion of looking for planets. Nonetheless, Baglin, the planet hunter in spite of herself, beat Borucki into space with it. Now he is nipping at her heels. Kepler’s bureaucratic history was even more tortured than Corot’s, but the spacecraft is headed for launch on April 10, 2009, and astronomers are counting on it to settle the question of just how common Earths are—a result that will guide the whole future search for life in the universe.
Borucki well remembers the effect he had with that 1984 paper, published in the journal Icarus. “There was no effect,” he says. “It was pretty much ignored.” At that time most researchers thought the way to look for other planets was through astrometry. Borucki was convinced that looking for planetary transits through photometry would be simpler and cheaper. Measuring the brightness of a star over time, he reasoned, would require a much smaller space telescope than trying to take a picture sharp enough to resolve a planet or a tiny loop in the star’s trajectory.
Borucki’s peers were skeptical, though—first, that a transit of an Earth would even be distinguishable against the background noise of the star’s fluctuating light, and second, that it was possible to monitor 5,000 sunlike stars at once, as he proposed to do. Research on the sun during the 1980s laid the first objection to rest; starlight turned out to be a lot less noisy than astronomers had thought. Over periods of days, which is how long it would take an Earth to pass in front of it, the sun’s output varies by only around 10 parts per million, whereas the passage of an Earth would dim it by 84 parts per million.
So such a transit was in principle detectable, but Borucki’s initial idea for a detector still raised eyebrows. He wanted to drill 5,000 holes, one for each star, into a metal template and put it near the focal plane of the telescope, with an individual photodiode and integrated circuit behind each hole. “People in the industry refused to even talk to us about that design,” recalls David Koch, the deputy principal investigator, whom Borucki roped in to his quest in 1992. Borucki’s basic concept was rescued by the emergence of the charge-coupled device, or CCD—the light-sensing chip that was then new and is now in hundreds of millions of digital cameras. A CCD can record the brightness of many stars at once, thus eliminating the need for thousands of photodiodes.
Borucki and Koch first proposed their mission to NASA in 1994; they called it FRESIP, for “Frequency of Earth-Size Inner Planets.” They proposed it again in 1996, 1998, and 2000. “Each time they came back with a list of reasons why we weren’t selected,” Koch says. “It won’t work because of this, it won’t work because of that, they said. And we went back and worked on it until we eliminated every reason they couldn’t select us.” The researchers made sure, for instance, that the minute vibrations of the spinning gyroscopes that kept the telescope pointed at its target would not drown out the signal from a planet. The name of the spacecraft was an easier sell. After the first proposal, Koch suggested naming it Kepler, after the discoverer of the laws of planetary motion. In 2001 NASA finally approved the mission. Over in Europe Corot was getting the go-ahead at around the same time, albeit at a much lower budget.
Though the French won the race into orbit, Kepler will have a telescope measuring 95 centimeters (37.4 inches), 3.5 times the diameter of Corot’s, with a field of view more than 10 times as large. Above all, it will be in orbit around the sun, trailing behind Earth, whereas Corot is in low Earth orbit over the poles. To avoid looking too close either to the sun or to Earth, Corot has to turn 180 degrees twice a year, which is why it can’t look longer than five months at one set of stars. As it orbits the sun, Kepler will stare at the same spot for its entire mission—which will allow it to detect true Earth analogues, those circling stars like the sun in orbits lasting about a year. “The French will find the first terrestrial-size planet,” Borucki predicts. “But those planets won’t be in the habitable zone. We’ll find the first one in the habitable zone.”
Actually, he expects to find several hundred. Kepler’s field of view covers 100 square degrees, around 0.25 percent of the sky, or as Koch puts it, “about two scoops of the Big Dipper.” Koch chose the most star-rich patch he could find in the U.S. Naval Observatory star catalog. It lies just out of the plane of the Milky Way, between the bright stars Vega and Deneb. Of the 13 million stars cataloged in that field of view, Kepler will monitor a subset that are most like the sun in size, mass, and age—stars that are quiet and sedate like our own. Its 42 CCD detectors lie just outside the telescope’s focus. “You don’t have to have a sharp image,” Borucki explains. “In fact, a fuzzy image is better”—the starlight is spread over more pixels, which keeps them from saturating as quickly. Because it has CCDs, Kepler will be able to monitor 100,000 stars simultaneously. It will spot variations as small as 10 parts per million in their light output.
When Kepler lifts off from Cape Canaveral on April 10, Borucki will have spent 17 years on his idea. “You don’t spend this long and then quit 90 percent of the way through,” he says. “I want to see the answers. I want to hold the data in my hand. I want to write the paper that says, ‘This is how many Earths there are.’” He may find none—which would be the most surprising result of all. Neither Borucki nor anyone in the planet-hunting business expects or wants that. It would mean Earths are extremely rare at best. It would mean we really might be alone in the galaxy, if not the universe.
By the time Kepler reports its first results, the exoplanet landscape will no doubt have changed yet again. Astronomers all over the world are hunting a second Earth. In May, Drake Deming of NASA was collecting data he hoped might reveal a super-Earth in the habitable zone of a red dwarf (a small and relatively cool star) called Gliese 436; NASA had allowed him to use a spacecraft called Epoxi, which is on its way to a rendezvous with a comet, to observe several stars that are already known to have planets. Also, last May Debra Fischer of San Francisco State University began training a small telescope in Chile on Alpha Centauri, a pair of sunlike stars that are the closest ones to Earth. Following up on an idea of Greg Laughlin’s, she believes that by observing it steadily for several years, she might detect not only an Earth but a Mercury and a Venus as well. “That’s what we’re going for—not just the first Earth but the first terrestrial planetary system,” she says. “It’s a pretty bold and crazy plan.”
Although Kepler and Corot are focusing on sunlike stars that could support true analogues of Earth, much of the action at ground-based telescopes is concentrating on red dwarf stars, for the simple reason that planets are easier to find there. An Earth-like planet would cause a bigger wobble and a darker transit in a red dwarf than in a sun, and the effect would be even more pronounced if the planet were in the habitable zone—because the habitable zone, where liquid water can exist, lies closer to a cool red dwarf. In the fall of 2007 David Charbonneau of Harvard began deploying a network of small telescopes in Arizona that will be focused on detecting transiting super-Earths in the habitable zones of red dwarf stars. The Swiss and California teams think they can do the same with the wobble technique. Of the super-Earths they’ve discovered so far, some—including the one around Gliese 876—orbit red dwarfs, though none lie in the habitable zone.
No one knows for sure whether a rocky planet in a red dwarf’s habitable zone would truly be habitable. Until recently, in fact, astronomers assumed it wouldn’t be. A planet so close to a star might have been blasted and sterilized, early in its existence, by flares from the star. It might be gravitationally locked, with one face always pointing toward the star, the way the moon always points the same face toward Earth. In that case the whole atmosphere might have frozen and snowed out on the dark side, leaving the planet airless and barren. But after having some of their preconceptions shattered by the discovery of Jupiter-size planets orbiting their stars in less than two days, planet hunters are no longer so confident of the others. Life might emerge on a red dwarf planet, some now think, after the star has aged and its flares have settled down; winds on the planet might transport heat from one hemisphere to the other, keeping the atmosphere from freezing. After a workshop on red dwarfs in 2005, Jill Tarter of the SETI Institute—a leading thinker on alien life—and her colleagues published an analysis that convinced many researchers that red dwarfs are worthy targets for Earth hunters. That’s a happy conclusion, given that red dwarfs are the most common stars in the galaxy and also the easiest targets for ground-based telescopes.
But even if a habitable Earth-like world is found first from the ground, it will most likely take a space observatory to search for the chemical signals that tell us what we really want to know: Is anything living out there? If the planet is one that can be observed transiting, it just might be possible to provide a hint of an answer in the next few years. As a transiting planet passes in front of its star, some starlight passes through the planet’s atmosphere and continues on toward Earth—minus certain spectral frequencies that have been absorbed by molecules in the atmosphere. In 2001, using Hubble, Charbonneau and his colleagues detected the first exoplanetary atmosphere that way; it belonged to a hot Jupiter called HD 209458 b, and it contained sodium, they said. Three years later Charbonneau found himself locked in a race with Deming to be the first to detect the flip side of a planetary transit—the moment, called secondary eclipse, when a planet passes behind its star. This time it was Deming who was observing HD 209458 b, with the Spitzer Space Telescope, an orbiting infrared observatory. Charbonneau, he knew, had collected data on a different hot Jupiter a month earlier. “We didn’t want to be second,” Deming recalls. “I was analyzing data while I was eating Christmas dinner. I had to catch Dave.” In the end they published papers simultaneously and held a joint press conference.
What each had done for the first time was detect an exoplanet’s photons. No telescope yet can spatially distinguish an exoplanet from its star; the distance between them is too small and the brightness contrast too large. A Jupiter adds about a billionth to the visible light of a sunlike star, and about a ten-thousandth to the star’s infrared glow (planets give off more heat than they do reflected starlight). By observing the combined infrared radiation of star and planet with Spitzer and then subtracting the radiation recorded from the star alone when it hid the planet, Deming and Charbonneau had detected the heat of the planet itself. From that they could calculate its temperature; Charbonneau’s team has since been able to create a crude weather map of their exoplanet, HD 189733 b, which showed that fierce winds must be spreading heat around its surface. Others using the secondary eclipse technique have detected evidence for water vapor and methane in the atmosphere of HD 189733 b.
These findings are trials for the far tougher task of picking apart the light of an Earth-like planet, much smaller and farther from its star (and thus far dimmer) than the hot Jupiters studied to date. Spitzer wasn’t designed to measure the spectrum of hot Jupiters, but it did. And the James Webb Space Telescope, which is slated to replace Spitzer in 2013, has not been designed to detect the spectrum of Earth-like exoplanets—but with its 6.5-meter (21.3-foot) mirror, nearly eight times the diameter of Spitzer’s, Deming and Charbonneau think it might. Other astronomers are more cautious. “I certainly wouldn’t claim JWST is going to prove habitability, because it’s not,” says Mark Clampin, project scientist for the observatory.
Sara Seager of MIT, who collaborates with Deming, is trying to figure out which spectral signatures in a planet’s atmosphere would provide the best evidence for signs of life. Water vapor is indicative of liquid surface water, which is necessary but not sufficient for life as we know it. Oxygen, which would quickly react out of Earth’s atmosphere if it weren’t continually produced by plants, is closer to a smoking gun, especially if it were seen together with methane. Then there is what Seager has dubbed “vegetation’s red edge”: At wavelengths of 700 to 750 nanometers, at the red end of the visible range, the reflectance of leafy green plants sharply increases, to four or five times what it is even at green wavelengths.
Whether the Webb can detect such a signature is not yet known, and even if it can, the data probably won’t be definitive. Deming and Charbonneau’s secondary eclipse technique, ingenious as it is, lacks the power to distinguish between life and something else. Making that distinction will require a new kind of space telescope. That’s where Corot and especially Kepler come in. They won’t provide targets for that future space telescope, unfortunately. To monitor many stars and maximize its chances of finding Earths, Kepler is forced to monitor distant ones; any Earths it finds will most likely be about 300 light-years away, too far for any currently imaginable space telescopes to take a spectrum from. What Kepler will do is tell astronomers—and NASA and ESA—what sort of space telescope it will take. If nearly every sunlike star has an Earth, we might find life around a relatively nearby star, with a relatively small telescope. If Earths are rare, the next telescope will have to be big.
Eventually such a telescope will get built—and if the pace of discovery remains as rapid as it is now, that day will come sooner rather than later. Finding convincing evidence for extraterrestrial life may take decades, but that is not a long time given the stakes. “Throughout recorded history we’ve had this question: Are we alone?” Tarter says. “For millennia, all we could do was ask the philosophers. Suddenly we have a way of looking for an answer that is not based on a belief system. I live in the first generation of humans that is able to do that. I think that’s extremely exciting.”
Other astronomers too feel acutely the historic nature of their quest. One of the last things to be mounted on Kepler this fall, before it makes the journey to Canaveral for the launch, will be a metal plaque engraved with the names of all 2,000 scientists, engineers, and managers who contributed to its mission, which will run through 2012. Kepler will follow a 53-week orbit around the sun, meaning that it will steadily drift farther behind Earth. “It loses a week a year,” Borucki says. “So 53 years after launch, it will come back to Earth. At that point, I expect, people will go up and pick up the spacecraft and put it in the Smithsonian. I know that sounds far-fetched. But I really think it will happen.”
