Chapter 8

Alien Earths

In Search of Wet, Rocky Habitats

The brilliant floodlights came on at dusk, with their crisscrossing beams pointed at the shiny Delta-II rocket perched on Pad 17-B at Cape Canaveral, Florida. The sky above was perfectly clear, with the Moon more than half full. The flawless liftoff occurred at 10:49 p.m. on March 6, 2009, as hundreds watched from within the Kennedy Space Center and hundreds more from the beach at Jetty Park a few kilometers away. One man among the select crowd within the compound, Bill Borucki, had waited longer, and fought harder, than anybody else to be there. The rocket’s precious cargo—a satellite observatory named Kepler, designed to search for transiting Earth twins around distant stars—was born out of Borucki’s dogged determination in the face of repeated setbacks and obstacles for a quarter century. “The night launch was spectacular,” Borucki told me. “At that moment, I thought of the hundreds of people who helped build Kepler. I felt as if their spirits were rising into the sky with it.”

Kepler offers the best chance yet of finding an Earth-size planet in an Earth-like orbit around a Sun-like star—a world that is suitable for life as we know it. Over its three-and-a-half-year lifetime, Kepler is expected to reveal dozens if not hundreds of candidates. The challenge lies in confirming them independently with radial velocity measurements, because most of its target stars are too faint to yield high-quality spectra of, even with the largest current telescopes. But the Kepler mission will certainly give us a better sense of how common Earth-like planets are in the Galaxy. Meanwhile, other surveys, using ground-based instruments, are foraging through the solar neighborhood for habitable worlds. Some surveys target red dwarf stars, not only because it would be easier to espy small planets in their midst but also because such stars are the most common variety. Already both radial velocity surveys and ground-based transit searches have turned up planets only a few times heftier than the Earth. Scientists are debating whether such “super-Earths” would be hospitable for life.

The Long Road

Borucki had first envisioned a planet-hunting telescope in a 1984 paper with Audrey Summers. But the seeds of his quest—some might call it obsession—were sown much earlier, during a boyhood spent assembling and launching rockets with his friends in rural Wisconsin. The only job he applied for, after completing a master’s degree in physics at the University of Wisconsin at Madison, was at the NASA Ames Research Center in California. He arrived there in 1962, a year after President John F. Kennedy had committed the United States to landing a man on the Moon. His first assignment was to help design the heat shields for Apollo. He developed instruments to measure the radiation from sample materials accelerated to high speeds in the lab. As the Apollo program came to a successful completion, Borucki started working on theoretical models of the Earth’s atmosphere to investigate the impact of releasing various chemicals into it. He also conducted laboratory experiments using lasers in flasks to explore the effects of lightning on planetary atmospheres and looked for lightning on Venus and Jupiter.

By the early 1980s, inspired by discussions with colleagues and workshops he had attended at Ames, Borucki’s attention turned to planets around other stars, a far-out topic at the time. He was intrigued by the possibility of detecting them through transits. Frank Rosenblatt, a Cornell computer scientist widely regarded as a pioneer of neural networks, had proposed the method in 1971, but he died in a boating accident soon after. Borucki not only developed the idea further in his own papers but also set out to identify detectors with sufficient precision to record the minuscule dips in a star’s brightness due to a transiting Earth-size planet. He initially focused on silicon diodes. However, many people were skeptical of their performance, especially in the unforgiving environment of space. So, in the late 1980s, Borucki’s team switched to charge-coupled devices, or CCDs, a technology more familiar to astronomers and one that had seen rapid progress. Throughout that period, the director of Ames supported Borucki’s fledgling enterprise with his discretionary fund. “People would say it is an impossible dream, that I was mistaken. But I was able to get enough funding to continue,” Borucki said.

By 1992, Borucki was ready to present his concept for a space telescope, dubbed FRESIP for Frequency of Earth-Sized Inner Planets, to the space agency. Two years later, he and his collaborators submitted a formal proposal to NASA’s Discovery Program, created to fund “faster, cheaper” missions that could launch within three years of selection for under 300 million dollars. The reviewers liked the science but didn’t trust the budget. Borucki’s team renamed the project Kepler and tried again at the next Discovery competition in 1996, with three independent verifications of their cost estimates. This time, there was a new objection: the CCDs would not be able to monitor tens of thousands of stars simultaneously with sufficient precision. So the team built a CCD camera, tested it at the nearby Lick Observatory, and came back for the third Discovery round two years later. Again, bad news: the spacecraft’s jitter could reduce the precision of measurements, the reviewers said. But this time NASA headquarters gave them a half-million dollars to build an end-to-end mock-up of the entire telescope system to demonstrate that it would work. Ames contributed another half million. Over the next six months, Borucki’s team built a 3-meter-tall contraption, with a CCD camera at one end and an artificial star field—light streaming through 1,600 tiny laser-carved holes in a stainless-steel plate—at the other. They tried to mimic all possible sources of “noise,” including spacecraft jitter, and still managed to detect simulated planetary transits.

“Year after year, we met each objection with studies that confirmed Kepler’s feasibility, we came back with evidence that it will work” Borucki said. Finally it made the cut in December 2001, perhaps because the reviewers had run out of criticisms or because the discovery of a transiting Jupiter from the ground a year earlier (see chapter 5) had vindicated the technique. By then, plans for a competing European mission, dubbed Eddington, were also in the works. The European Space Agency eventually canceled Eddington due to budget problems, while Kepler went ahead. For Borucki, the ultimate goal more than justifies the long and twisted road he has had to endure. “If we want to explore whether there are other civilizations out there, the first step is to find planets that are habitable,” he explained. “That’s what Kepler is designed to do.”

Goldilocks Planets

What makes a world habitable? Being the right size is probably the first requirement. If a planet were more than about ten times the Earth’s mass, it would accrete a huge atmosphere and become a gas giant like Jupiter or Saturn with no solid surface. Any living organisms trying to form in their dense atmospheres would be carried alternately to frozen heights and overheated depths by the strong convection cells, making survival difficult. Thus, despite the detection of simple amino acids in Jupiter’s atmosphere, scientists think gas giants are unlikely to make safe harbors for life. At the other end of the scale, if a planet is too puny, that spells trouble too. It wouldn’t be able to hold onto a substantial atmosphere, so the oceans would boil off. What’s more, a small planet would not have a stable climate over billions of years. The reason has to do with plate tectonics, which involves the movement of large chunks of a planet’s crust. Where these plates run into each other, mountain ranges—like the Himalayas on Earth—build up. The process also enables complex chemistry and long-term recycling of substances like carbon dioxide among the atmosphere, the ocean, and the crust, making a stable climate possible. A rocky planet would have to be at least one-third as massive as the Earth to enable plate tectonics. Mars falls below that threshold.

Being in a fairly circular orbit is also essential, to ensure even heating from the parent star during the course of a year. A planet in a highly elongated orbit would suffer from extreme swings in temperature as its distance from the star varies.

Beyond those basic requirements, it is all about location, location, location. Being located around the right kind of star is a must. What makes a good parent star? It should live long enough for life to have sufficient time to develop and evolve. Massive stars live fast and die young; they exhaust their hydrogen fuel supply within tens of hundreds of millions of years.¹ Low-mass red dwarfs, on the other hand, have incredibly long lives—hundreds of billions of years. With a lifetime of 10 billion years, the Sun falls somewhere in the middle of the range. As in real estate, neighbors matter, too. If the parent star has a companion, as many stars do, it should be either very close so that a planet can have a stable orbit encircling both—imagine the spectacular double sunsets!—or very far so that a planet around one star isn’t affected much by the other. Luckily, most binary star systems appear to be safe. Overcrowded neighborhoods may be hazardous to life. Within dense star clusters, for example, close encounters between stellar neighbors could disrupt each other’s planetary systems. The Sun probably formed in a loose group of stars that has since dispersed. Today, stars in our neighborhood are safely separated, by a few light-years on average. The parent star’s location in the Galaxy might matter, too. Stars closer to the center move faster than the Galaxy’s spiral pattern, so their planetary retinue may suffer as they run into dense molecular clouds or clusters of massive newborn stars concentrated in the spiral arms. Again, luckily for us, the Sun is located in the outer suburbs of the Galaxy, where its rotation speed is roughly the same as that of the spiral arms, so it can remain safely suspended between two arms.

Astronomers usually define a habitable planet as one that can sustain liquid water on its surface.² That means the planet must not be too close to its star or too far from it. The not-too-hot, not-too-cold “Goldilocks” region around a star, where the temperature is just right for liquid water, is called the habitable zone. The Earth is safely within the Sun’s habitable zone. Venus, about 30 percent closer in, is not: the scorching heat probably boiled off its surface water early on. Mars, about 50 percent farther from the Sun than the Earth, is much colder but still barely inside the present-day habitable zone. In fact, Mars appears to have had flowing water and large lakes in the past. “The problem is that Mars is too small to recycle carbonates,” explained Jim Kasting, a planetary scientist at Pennsylvania State University. “An Earth-mass planet at Mars distance would be habitable, because carbon dioxide would accumulate in its atmosphere, warming the planet through the greenhouse effect.”

Figure 8.1. The location of the habitable zone, where liquid water can exist, depends on the type of star.

Kasting grew up in Huntsville, Alabama, home to NASA’s Marshall Space Flight Center. That’s where the gigantic Saturn V rockets were built in the 1960s to launch the Apollo astronauts to the Moon. “The whole city would shake when they tested various stages of the Saturn V,” he reminisced. The head of NASA’s rocket program, Wernher von Braun, a German engineer who had designed the deadly V-2 combat rocket during the Second World War and surrendered to the Americans in 1945 with many of his staff, was an early hero. As a teenager, Kasting was also captivated by science fiction written by the likes of Robert Heinlein and Isaac Asimov. Kasting studied physics and chemistry as an undergraduate at Harvard. His PhD thesis in atmospheric science at the University of Michigan focused on the rise of oxygen in the Earth’s atmosphere. First, he needed to calculate models of the atmosphere before the appearance of photosynthesis. He found that when the Earth was cold, carbon dioxide would build up. But the rising CO₂ levels, in turn, would cause greenhouse warming. The early Earth might have switched between frozen and wet states as a result of this CO₂ climate feedback, he concluded. Kasting’s attention turned to questions of planetary habitability in the early 1990s. One issue he and his colleagues have studied is the length of time for which Earth has been habitable. The problem is that the Sun was about 30 percent fainter4.6 billion years ago than it is today. It has been getting gradually brighter since. As Carl Sagan and George Mullen, both at Cornell University at the time, first pointed out in 1972, the implications for the early Earth’s climate are dramatic. With much less incoming sunlight, our planet’s surface temperature would have been below freezing until about 2 billion years ago. However, the geological evidence shows that both liquid water and life were present much earlier, as far back as 3.5 billion years ago. In fact, the oldest zircons, minerals that appear to have formed in liquid water, date from even earlier, some 4.3 billion years ago. The solution to this “faint young Sun paradox” probably lies in higher concentrations of carbon dioxide and methane—both greenhouse gases—in the early Earth’s atmosphere, according to Kasting.³ As mentioned earlier, carbon is recycled between CO₂ in the atmosphere and carbonate minerals in the oceans and the crust, thanks to plate tectonics. This long-term cycle has a negative feedback loop: in other words, at higher temperatures more of the carbon is tied up in carbonates (reducing greenhouse warming), whereas at lower temperatures more of it ends up in atmospheric CO₂ (increasing greenhouse warming). The response time of this feedback loop is hundreds of thousands to millions of years. So it is too slow to counteract global warming caused by humans, but it is fast enough to stabilize the climate over billions of years. As a result, the Earth has been continuously habitable for a large fraction of its history.

Figure 8.2. The greenhouse effect. Sunlight at visible wavelengths heats the planet’s surface, but the planet surface re-radiates at infrared wavelengths, which are trapped by certain gases in the atmosphere.

Even today, a little greenhouse warming is essential for keeping the Earth warm enough for liquid water. On Venus, however, the greenhouse effect has gone awry, raising its surface temperature by some 500 degrees Celsius. That’s the result of CO₂ from volcanoes building up in its atmosphere, because the carbon cycle cannot operate without water, which was lost early on. The story is different on Mars. Farther from the Sun, it would have needed a stronger greenhouse effect to warm up the surface. Its small size meant an early end to geologic activity such as volcanism—thus no mechanism existed for recycling CO₂ to stabilize the climate. The thin atmosphere does not help, either. But Mars might have had periods of substantial greenhouse warming in the past, as the evidence of ancient lakes and riverbeds suggests.

Kasting and his colleagues found that habitable zones around solar-type stars are fairly wide, once they accounted for the stabilizing feedback from the carbon cycle. At present, the inner edge of the zone is just inside the Earth’s orbit, at about 0.95 AU; any closer to the Sun, water would boil off. The outer edge could be anywhere between 1.4 and 2.4 AU, depending on the amount of greenhouse warming. Stars a bit more massive than the Sun have somewhat wider habitable zones located farther out. But stars heftier than about two solar masses do not make great hosts because of their rather short lives.

Extreme Living

Over 80 percent of stars in the Galaxy are small, dim red dwarfs (also known as M stars). Because they are so numerous and have extremely long lives, the habitability of planets in their midst is a critical question. Initially, scientists assumed that red dwarf systems would make poor havens for life for two reasons. First, their habitable zones are not only narrow—about a tenth as wide as the Sun’s—but also so close in—much closer than Mercury is to the Sun—that any planets orbiting there would be tidally locked in. That means one side of the planet always faces the star, just like one side of the Moon constantly faces the Earth. You might expect the dayside to be scorching hot while the nightside remains in an eternal freeze, with a giant icecap gathering all of the planet’s moisture. Second, many of these stars are terribly tempestuous, especially when they are young, frequently putting out strong fares of harmful ultraviolet radiation and fast-moving particles. Any planets in the close-in habitable zone might be sterilized.

However, scientists now think that some red dwarfs could harbor habitable worlds after all. If a planet is massive enough to retain a substantial greenhouse atmosphere, wind circulation could keep the temperatures fairly mild all around the planet even if it is tidally locked. The incoming ultraviolet radiation would simply turn oxygen in the planet’s stratosphere into ozone, which in turn would shield the surface from harmful fares. Besides, some red dwarfs act up less than others. At least, these more quiescent ones could permit life to develop on Earth-size rocky planets with moderately dense atmospheres located in their narrow, close-in habitable zones. Creatures on such a planet would be living in a world of contrasts, with the day lasting forever on one half and a permanent night on the other. On the sunlit side, the parent star would cover a big chunk of the sky. From the dark side, one might see spectacular auroras during stellar fares when fast-moving particles enter the planet’s magnetic field.

There is another reason that astronomers are rooting for habitable rocky planets around red dwarfs: they would be easier to detect than those orbiting Sun-like stars. A given planet’s gravitational tug would have a bigger effect on a lower-mass star, rendering larger—thus easier to measure—velocity shifts in its spectral lines in Doppler observations. The transit method favors them immensely, too. First, habitable planets would be in tighter orbits around red dwarfs, thus increasing the likelihood that they will be seen in transit. Second, because red dwarfs are smaller, the relative dip in brightness when a planet passes in front would be bigger; thus a small rocky planet’s transit might be measurable even with a ground-based telescope.

Life could originate not only on Earth-size planets but also on big rocky moons of gas giants in a star’s habitable zone—as in the hypothetical moon called Pandora orbiting a gas giant planet depicted in the 2009 Hollywood blockbuster Avatar. In our own solar system, Saturn’s moon Titan has a dense atmosphere rich in methane. Tides between a moon and its planet could provide an additional source of heat, as is the case with Jupiter’s Io and Europa. In fact, thanks to radial velocity searches, we already know of giant planets in the habitable zones of several nearby stars, including mu Arae, HD 23079, and HD 28185. But we do not know yet whether any of these planets harbor big satellites.

Super-Earths

As the precision of Doppler measurements has improved, thanks to better instrumentation, it is now possible to detect velocity shifts as small as 1 meter per second—about the pace of a leisurely stroll—in nearby low-mass stars. That is enough to reveal the presence of planets only a few times the Earth’s mass. The newest denizens in the planetary zoo to be identified are two to ten times more massive than our planet. These strange new worlds, not seen in our solar system, are called super-Earths.

The first of its kind to be announced, back in 2005, is a planet with a minimum mass of 7.5 times the Earth’s, orbiting the nearby red dwarf star Gliese 876. The host star was already known to possess two gas giants. In fact, it was only after accounting for the resonant interactions between the two larger planets that astronomers were able to decipher the presence of a third, less-massive body. Jack Lissauer and Eugenio Rivera at the NASA Ames Research Center mined many years of data from the California-Carnegie planet search team, led by Geoffrey Marcy and Paul Butler, to uncover the super-Earth in a two-day orbit. The planet is so close to the star that it is probably tidally locked, taking just as long to rotate on its own axis as to revolve around the star. Its permanent dayside, constantly baked in starlight, is likely oven-hot, at 200–400 degrees Celsius. The nightside temperature would depend on how efficient its atmosphere is at distributing the heat through wind streams.

Since then, Doppler surveys have identified numerous super-Earths not only around red dwarfs but also around stars similar to the Sun. Three of them, found by the Geneva team, circle the same star, HD 40307, in orbits much tighter than Mercury’s. “Clearly these planets are only the tip of the iceberg,” Mayor said at the time of their announcement in 2008. He claimed that a third of all Sun-like stars in their sample betray hints of super-Earths or Neptune-mass planets in orbits shorter than fifty days.

The coolest, and the most distant, super-Earth yet was discovered through gravitational microlensing, when a nearby star and its planet magnified the light of a more distant star temporarily (see chapter 5). The 5.5-Earth-mass planet roughly 2.5 AU from a dwarf star some 20,000 light-years away was announced in 2006 by a large team of astronomers led by Jean-Philippe Beaulieu at the Institute of Astrophysics in Paris. While the Doppler and transit techniques are best suited to finding planets close to their star, microlensing favors the detection of super-Earths farther out. This particular planet—dubbed OGLE 2005-BLG-390Lb—is almost certainly in a deep freeze, with temperatures approaching that of Pluto, given its distance from the faint parent star.

The least massive planet yet found and the only super-Earth reported to be located in a habitable zone both circle a red dwarf called Gliese 581 just twenty light-years away. In fact, this planetary system may contain six low-mass planets in total. Michel Mayor’s Geneva-based team found the first four planets between 2005 and 2009, using the HARPS instrument. The innermost and smallest of them, Gliese 581e, is a 2-Earth-mass body circling the star every three days. While its surface is way too hot for liquid water, as the record holder for the least massive planet yet, it is likely rocky, and its discovery is a key step in the path toward Earth twins elsewhere. Two of the other planets in the system straddle the habitable zone, but are probably either too hot or too cold to harbor liquid water on the surface.

Table 8.1 Some Super-Earths Orbiting Normal Stars

There was a furry of press attention in September 2010, when a team led by Steve Vogt at UC Santa Cruz and Paul Butler of the Carnegie Institution of Washington announced two more planets around Gliese 581, one of them well within the habitable zone. Gliese 581g, as it is dubbed, was reported to orbit the star every 37 days, and would be tidally locked, with one side basking in perpetual sunshine while the other side remains in eternal darkness. “This is the first exoplanet that has the right conditions for water to exist on its surface,” Vogt told the New York Times. “This is really the first Goldilocks planet,” Butler added.

However, at an exoplanet conference in Turin, Italy, in October 2010, which I attended, the Swiss team cast doubt on the two new planets. A Facebook post I made from the conference about the Swiss team’s presentation was picked up by several bloggers and journalists. “If a signal corresponding to the announced Gliese 581g planet was present in our [HARPS] data, we should have been able to detect it,” Francesco Pepe of the Geneva Observatory told Science News. “From these data we easily recover the four previously announced planets. However, we do not see any evidence for a fifth planet in an orbit of 37 days,” he added. Meanwhile, Steve Vogt stood by their team’s findings. So, for now, claims of a low-mass planet in the habitable zone of Gliese 581 remain contested. As I told Science News, “Given the extremely interesting implications of such a discovery, it’s important to have independent confirmation.”

Even though only a handful of super-Earths have been identified so far, many researchers think they are more common than gas giants. That is because it should be easier to build up small rocky planets in protoplanetary disks around young stars. Besides, red dwarfs are the most common type of star in the Galaxy, and their low-mass disks favor the formation of lighter planets. In the widely accepted core-accretion model, solid particles coagulate into planetary cores, which then accumulate gas from the surrounding disk if they grow massive enough and the gas has not dispersed by then. The threshold for accreting and retaining a hydrogen envelope is a bit uncertain, but theorists estimate it to be about 10 Earth masses. That sets an upper limit on how massive a super-Earth can grow before turning into a cousin of Neptune. If super-Earths are common, why isn’t there one in our solar system? It is probably a matter of coincidence, according to Dimitar Sasselov of Harvard University. “But,” he added, “it is possible that Jupiter prevented the formation of a more massive terrestrial planet in our case.”

Rock or Ice?

Super-Earths probably come in two main favors: ocean planets, in which water (or ice) makes up over a tenth of the mass, and rocky Earth-like planets, which could still host oceans (as the Earth does, even though water accounts for only 0.05 percent of its mass). Both types would contain solids, such as silicates and metals; volatiles, such as water and ammonia; and trace amounts of hydrogen and noble gases. The heavier compounds, like iron alloys, quickly settle into a core, while the volatiles precipitate above a silicate mantle. If there is a lot of water, most of it is likely to be under high pressure, in the form of ice even at high temperatures. Since ice is less dense than silicates, an ocean planet would be bigger than a rocky planet of the same mass. If astronomers could measure the mass of a super-Earth to an accuracy of 10 percent and its radius to 5 percent, they could probably tell whether it is a rocky or ocean planet, according to calculations by Sasselov and his colleagues at Harvard, Diana Valencia and Richard O’Connell.

Figure 8.3. For the same mass, an icy planet (left) would be bigger than a rocky one (right).

The first super-Earth to be caught in transit is a strange one called COROT-7b, found by the French COROT satellite. It is one of the smallest extrasolar worlds to have its radius measured, at less than twice that of the Earth. Its orbital period of twenty hours implies extreme proximity to the host star, some twenty times closer in than Mercury is to the Sun. The parent star is a bit smaller and younger than the Sun and is located nearly 400 light-years away. Stellar activity confounded the astronomers’ initial attempts to measure the planet’s mass using Doppler observations. But after careful monitoring with the HARPS instrument, the Geneva team was able to determine a mass about five times that of the Earth, making it one of the lightest extrasolar planets known. It is also the first alien world with strong evidence for a rocky composition, given a density similar to the Earth’s. With its star-hugging orbit, scientists estimate a scorching surface temperature of 1500 degrees Celsius on the dayside of COROT-7b. “The place may well look like Dante’s Inferno,” commented Didier Queloz of the Geneva Observatory. The planet’s surface may be covered in lava or boiling oceans, thus it is not a hospitable environment for life. Some scientists speculate that it could even be the remnant core of a Saturn or a Neptune whose atmosphere evaporated as it approached the star. The Doppler measurements also revealed a second super-Earth a bit farther out around the same star. The 8-Earth-mass outer world orbits the star in less than four days but does not pass in front of it, unlike the inner planet.

Astronomers cannot measure the radii of those super-Earths that do not transit in front of their star. Without knowing the radius, it is not possible to infer the bulk composition reliably. Some of them could be rocky, while others might be remnants of Neptune-like planets, stripped of their outer layers. It is likely that the super-Earths around Gliese 876, Gliese 581, and HD 40307 all formed beyond the “snow line” of their star systems—where water freezes and accretes easily into planetary cores—and migrated inward later. As they approached their stars, some of the surface ice may have melted. The OGLE 2005-BLG-390Lb planet remains stuck in the frozen outskirts.

Whether these particular planets are habitable or not, super-Earths found in the future may be excellent targets to search for biosignatures. While most researchers focus on Earth-like planets as potential habitats, super-Earths may in fact offer better odds, according to Harvard’s Sasselov. “In my opinion, super-Earths are the best places to look for life. Actually, they have higher habitability potential than Earth-mass planets,” he said. Calculations by Valencia, Sasselov, and O’Connell suggest that super-Earths would likely harbor thinner plates under more stress, fostering more vigorous geologic activity and, in turn, more efficient and possibly faster recycling of nutrients between the planet’s exterior and the interior. In fact, the Earth may be barely above the threshold: slightly less massive Venus is tectonically inactive. “Bigger is better when it comes to the habitability of rocky planets,” concluded Valencia.

Penn State’s Kasting is not so sure. He points out that a super-Earth would take longer to build up oxygen in its atmosphere through photosynthesis. “Therefore, my guess is that complex life is less likely to develop on a 2-Earth-mass planet by 5 billion years,” he said.

Exotic Mixtures

Terrestrial planets in our solar system are made mostly of silicon-oxygen compounds called silicates, in a variety of forms such as quartz and feldspar. But some low-and intermediate-mass planets around other stars, including super-Earths, may form substantially from carbon compounds and could contain layers of diamond. That’s the intriguing suggestion by Marc Kuchner of NASA’s Goddard Space Flight Center and Sara Seager of the Massachusetts Institute of Technology.

If a protoplanetary disk around a young star contained too much carbon or too little oxygen, carbon compounds like carbides (a hardy ceramic) and graphite (found in pencil lead) can condense out of the gas, instead of silicates. Under high pressure, graphite would turn into diamond, possibly forming layers of it many miles thick. Carbon planets could have iron cores, tar-covered surfaces, and atmospheres rich in hydrocarbons and carbon monoxide. “Such planets are likely to be rare, because you need an overabundance of carbon relative to oxygen in the disk for them to form,” said Harvard’s Sasselov.

Kuchner points out that carbon planets could form in much the same way as certain meteorites, the so-called enstatite chondrites, in our solar system. “The chemistry of some meteorites suggests that they could have formed from large quantities of carbon-rich dust. Some meteorites even contain tiny diamonds,” he explained. “All you have to do is to imagine the same mix on a planet scale.” Of course, that didn’t happen around the Sun. But Kuchner’s case may have received a boost from the recent finding that the disk around the 12-million-year-old star beta Pictoris is particularly carbon-rich. Based on observations with the Far Ultraviolet Spectroscopic Explorer satellite, a team led by Kuchner’s Goddard colleague Aki Roberge estimated that the beta Pic disk has nine times as much carbon as oxygen. That is quite a contrast to the Sun, which contains half as much carbon as oxygen. Now the question is whether beta Pic represents a different kind of planetary system—one where carbon planets might form—or whether it is just a phase that all planetary systems go through before the “excess” carbon, possibly released from asteroids and comets, is swept away by strong stellar winds.

Getting There

The race to find the first “second Earth” is now under way. The High Accuracy Radial-velocity Planet Search (HARPS) spectrograph on the European Southern Observatory’s 3.6-meter telescope at La Silla, Chile, used by the Geneva-led team of planet hunters, can detect Doppler shifts below 1 meter per second. To achieve such precision, the instrument’s temperature needs to be kept constant to within one-hundredth of a degree, so that tiny expansions or contractions of its components do not degrade the measurements. But that is still insufficient to detect the extremely subtle wobble caused by an Earth twin orbiting a Sun-like star. Such a planet would induce velocity shifts of only 0.1 meters per second. What’s more, the star would have to be tracked for an entire Earth-year to confirm its existence.

On the other hand, an Earth-mass planet in the habitable zone of a red dwarf might be almost within reach: the lower-mass star would show a bigger velocity shift, and the closer-in planet would complete an orbit in tens (rather than hundreds) of Earth-days. The Geneva team, with Michel Mayor at the helm, is now targeting about a hundred nearby red dwarfs with HARPS in their quest for terrestrial worlds. They are also observing hundreds of Sun-like stars to search for more super-Earths.

The 2.4-meter Automated Planet Finder telescope under construction at the Lick Observatory on Mount Hamilton, California, will also have a spectrograph designed to reach similar precisions. That project suffered from delays, in part due to a break-up of the California-Carnegie planet search team in 2007. Steve Vogt at the University of California, Santa Cruz, who designed the spectrograph, and Paul Butler of the Carnegie Institution of Washington have formed their own team, separate from Geoff Marcy at Berkeley and Debra Fischer, now at Yale. “Keeping people together in a close collaboration for a long time is hard. . . . In some ways, it’s like a marriage. It took about a year to recover from the team’s break-up,” Fischer explained. “Relations had been rocky for some time, but I wish we had worked things out. I have no hard feelings though,” she added. There is now a memorandum of understanding between the two groups to share time on the telescope, once it is completed.

Meanwhile, Fischer and her collaborator, Greg Laughlin of the University of California at Santa Cruz, are using a modest-size telescope on Cerro Tololo in the Chilean Andes to hone in on one target. At a mere four light-years away, alpha Centauri is the nearest star system to us. It consists of three stars altogether: alpha Centauri A and B are similar to the Sun and orbit each other every eighty years while the third, called Proxima Centauri, is a red dwarf in a much wider orbit. It’s the inner pair that draws the researchers’ interest. “We want to beat down the noise with a hundred thousand measurements and look for terrestrial planets in the habitable zones of these stars,” Fischer said. Laughlin’s theoretical work had shown that planets could exist in stable orbits within 2 AU of either star, although not everyone agrees. Other simulations suggest that planets are less likely to form in such a binary system. Fischer thinks the high-risk gamble is worth taking. “These stars are bright and close, they are metal-rich like many planet hosts, and their orbital inclination is favorable,” she explained. “And could you imagine the enormous public interest if we were to succeed?” Those who watched the movie Avatar will recognize alpha Centauri as the home of the supposedly inhabited moon Pandora. The alpha Centauri pair is also on the target list of the Geneva team, observing from the nearby La Silla observatory, though they do not spend as much time as the Fischer and Laughlin duo, preferring to survey a large sample of stars rather than gamble on one system in particular, however appealing it might be for terrestrial planet searches.

Ultimately, the stars themselves will set the detection limits for Doppler surveys. Starspots and other active regions contribute “noise” to velocity measurements. One solution might be to do Doppler surveys at near-infrared wavelengths, where the contrast between spots and rest of the stellar surface is less pronounced. Moreover, red dwarfs shine brighter in the near-infrared. One potential challenge, compared with the optical regime, is contamination by numerous absorption lines due to the Earth’s own atmosphere. However, simulations suggest that masking out the portions of the spectrum with the worst contamination would still leave enough wavelength coverage for precise Doppler measurements. Unfortunately, a plan to build a high-precision infrared spectrograph for the 8-meter Gemini telescope in Hawaii, to search for terrestrial planets around three hundred nearby red dwarfs, was canceled recently due to budget problems.

A transiting Earth-size planet occults only 1/10,000 of a Sun-like star, resulting in a brightness dip comparable to the dimming one might see if an insect were to crawl across a car headlight seen from several kilometers away. That effect is too subtle to detect from the ground. But a similar planet would cover a larger fraction—roughly one-thousandth—of a smaller red dwarf, so the corresponding transit signal is easier to measure. However, since red dwarfs are intrinsically faint, only the nearest ones make suitable targets, and those are distributed all over the sky. The MEarth (pronounced “mirth”) project, led by David Charbonneau at Harvard, plans to target two thousand nearby red dwarfs in search of transiting terrestrial planets using eight 0.4-meter robotic telescopes, located on Mount Hopkins, Arizona. Built with off-the-shelf technology, these instruments are no bigger than the backyard telescopes owned by many amateur astronomers.

The MEarth team tasted early success in 2009 with the discovery of a transiting super-Earth around the red dwarf GJ 1214, a mere forty light-years from the Sun. Less than three times the Earth’s size, it is one of the smallest exoplanets known. Follow-up Doppler measurements put its mass at just under 7 Earth masses. With an estimated temperature of 200 degrees Celsius, it is also the coolest transiting planet yet, despite being as warm as an oven. The planet’s low density suggests a very different makeup from the Earth’s. Theorists suggest that it could be a mini-Neptune, a steam world, or a rocky planet surrounded by a puffy atmosphere of hydrogen.

Recent infrared observations of its atmosphere by my graduate student Bryce Croll, taken during several transits using the Canada-France-Hawaii Telescope, favor the mini-Neptune description: a rocky core surrounded by a large gaseous envelope rich in hydrogen and helium. While that is disappointing for prospects of life, it is remarkable that we are able to characterize the atmosphere of a super-Earth with a modest-size telescope on the ground.

There is another possible avenue for finding rocky planets from the ground. Just as nineteenth-century astronomers inferred the presence of Neptune from its gravitational influence on Uranus’s orbit, a lower-mass planet’s presence might be uncovered by its effect on a transiting Jupiter in the same planetary system. The second planet’s tug would cause transit times of the bigger planet to deviate slightly from exact periodicity, especially if the two are in resonant orbits. The effect of an Earth-mass planet on a Jupiter’s transit times could be as large as a few minutes over several months. Since hot-Jupiter transits can be timed to a precision of a few tens of seconds with ground-based telescopes, several teams are already on the hunt for terrestrial planets using this method. So far, there has not been a positive detection, but researchers have managed to rule out Earth-mass worlds in resonance with the inner giant planets in a few systems.

On the space front, the 572-million-dollar Kepler mission is expected to open the floodgates for discovering rocky worlds around other stars. “Kepler is the most exciting mission right now. It will revolutionize our view of planetary systems by telling us about the frequency of terrestrial planets,” said planet hunter Debra Fischer. “I personally expect it to find enormous numbers of low-mass planets,” she added. For three and a half years, the 95-centimeter telescope will stare at a patch of the sky about the size your fist would cover when held at arm’s length. This star field, toward the constellations of Cygnus and Lyra, contains millions of stars. The Kepler team has selected 150,000 of them as prime targets for planet hunting, because of their stellar properties. The target stars span the range from puny red dwarfs to massive B-type stars, but a significant number are fairly similar to the Sun. The spacecraft’s Earth-trailing orbit around the Sun will ensure virtually continuous monitoring of the Cygnus-Lyra field for the entire duration of the mission.

Four weeks after its 2009 March launch, with in-orbit checks complete, astronomers received the first image of the star field taken by Kepler. “That moment for me was even more profound than the launch,” team member Natalie Batalha of NASA’s Ames Research Center told me. “When I saw all these stars in nice crisp focus, it was quite something.” In early June, the first science data, consisting of ten days’ worth of observations, arrived. The dataset included observations of HAT-P-7, a previously known transiting hot Jupiter in a two-day orbit. Kepler’s measurements were precise enough to detect the changing phases of the planet as it circled its star. “This early result shows the detection system is performing right on the mark,” Batalha’s colleague David Koch told the media when the finding was released in early August. “It bodes well for Kepler’s prospects to be able to detect Earth-size planets.”

“We were overwhelmed by the number of new planet candidates in these commissioning data,” Batalha explained. “We started skimming the cream off the top.”

The goal was to get the most promising candidates to their collaborators for radial velocity follow-up at ground-based telescopes. Meanwhile, the team received thirty-three more days of data from Kepler. Not surprisingly, the satellite unraveled big planets in short-period orbits—hot Jupiters—first. At the January 2010 meeting of the American Astronomical Society, Borucki and Batalha presented the first five planets discovered by Kepler and confirmed by ground-based Doppler measurements. “There’s a deluge of data approaching,” Batalha said, looking ahead. “We are not going to confirm every candidate. We’ll throw over the fence more-massive planet candidates for the community to follow up and focus on the smallest candidates ourselves,” she added.

Based on results from earlier transit surveys from the ground, the Kepler team expects to find about 135 close-in giant planets and another thirty or so beyond the Earth-Sun orbital distance. Kepler is likely to find a few hundred super-Earth candidates as well. But the satellite observatory’s primary goal is to look for Earth-size planets in Earth-like orbits around solar-type stars. The mission lifetime is such that Kepler should be able to observe three consecutive transits of planets in one-year orbits. Assuming that all target stars have at least one terrestrial world in their habitable zone, the project scientists expect to find between fifty and 640 such candidates, depending on the size distribution of rocky planets. But these numbers are really just estimates based on uncertain assumptions. The actual numbers could be much higher or lower and will tell us how common or rare Earths are in the Galaxy. “We are all closet optimists,” Batalha admitted. “We expect Earths to be common, both for good scientific reasons plus that human inclination.” Getting reliable statistics on the planet population—how many planets of which kinds circle different types of stars—is Kepler’s strong suit. Bill Borucki has a sharper focus. “I want to know the frequency of Earth-like planets out there,” he told me. “Everything else we find is gravy, frosting on the cake.”

One critical challenge is disentangling real planetary transits from unrelated brightness dips—like those due to eclipsing double stars whose eclipses appear small because of blended light from a third star—that can mimic a planet’s signal. Perhaps one hundred of these candidates would be around stars bright enough for follow-up spectroscopic observations from large ground-based telescopes, to measure the star’s wobble due to the planet’s gravitational tug, and infer its mass. Borucki expects to marshal large amounts of telescope time for that enterprise with the help of his team members. Still, “we have to be as thrifty as possible with telescope time,” said Fischer, a veteran of Doppler surveys. “Some terrestrial planet candidates will be around stars bright enough for us to quickly confirm the minimum and maximum velocities of the wobble once we know their periods from Kepler,” she explained. “But longer-period planets in the habitable zone will be extremely difficult to confirm definitively, especially for solar-type stars. . . . The velocity amplitude goes way down for those.” However, given the huge implications of finding Earth-like worlds, she added, “you have to demonstrate at least a few.” Borucki also acknowledged, “confirming a small Earth-mass planet with radial velocity would be a huge challenge.”

Many of the transit candidate stars will be too faint for follow-up with even the largest telescopes on Earth.

“For those, people will lean more on the transits themselves,” said Fischer. “The concern is how high or low the false alarm rate will be,” Stephane Udry from the Geneva Observatory told me. “With precise photometry from space, it may be possible to rule out a lot of false positives and build confidence,” he added. In the end, many researchers expect Kepler to produce a handful of near Earth-twin detections and many more likely candidates. The latter’s confirmation may have to await the next generation of behemoths planned for the next decade, like the Thirty Meter Telescope and the European Extremely Large Telescope. In the meantime, Kepler is likely to give us a reliable statistical picture of habitable worlds in the Milky Way—extremely useful for guiding future missions that will image such planets, like the European Space Agency’s Darwin and NASA’s Terrestrial Planet Finder.

If Kepler finds a paucity of Earth-size planets in habitable zones, the implications will be profound. On the other hand, if it finds that terrestrial worlds are common, as many astronomers expect, that does not necessarily imply life is abundant in the cosmos. Venus and Earth are planetary twins in many ways: both are rocky worlds, and have about the same size, mass, density, and cloud-top temperature. Yet Venus, with a broiling surface temperature of 400 degrees Celsius, a crushing surface pressure ninety times that of the Earth’s at sea level, and a carbon dioxide atmosphere that confirms our worst fears of the greenhouse effect, cannot sustain life as we know it. That’s why finding, or even imaging, Earth-size planets elsewhere is one thing, but detecting life is quite another.

¹ Massive stars are crucial to life elsewhere, though. They produce key elements for life, including carbon, nitrogen, oxygen, calcium, and iron. These elements are scattered into space in the late stages of the massive stars’ lives and are later incorporated into molecular clouds, out of which new generations of stars and their planetary systems form.

² This definition might seem a bit limited, now that we know of living organisms in deep mines and under kilometers of ice (see chapter 9). Microbes might exist below the surface of Mars or the ice cover of Jupiter’s moon Europa. But pretty much the only way to detect such organisms is in situ, not through remote sensing from many light-years away. Therefore, when we discuss habitable worlds around other stars, it makes sense to search for those with surface water.

³ The rise of oxygen, some 2.3–2.7 billion years ago thanks to the evolution of photosynthetic microbes and plants, would have destroyed methane, removing an important greenhouse agent and probably causing the Earth’s first episode of global glaciation.