The two thousand stars in the N2K survey aren’t just bright; they’re also known to be rich in “metals.” This means something very different to astronomers than it does to anyone else. In astronomy jargon, a metal is any element other than hydrogen or helium, which is all of them. So while everyone agrees that iron and aluminum are metals, astronomers also include oxygen and carbon and nitrogen and neon in that category. Debra Fischer had been one of the first exoplaneteers to show that planets are more common orbiting metal-rich stars than metal-poor stars—not surprisingly, since planets are made in large part of these elements—so looking at these stars was a way to boost the odds of finding exoplanets. N2K uses giant telescopes in Hawaii and Chile to look for the radial velocities, and small, automated telescopes at the Fairborn Observatory in Arizona to look for transits.
And these were only a few of many transit searches that were under way. “When I attended a conference on extrasolar planets in Washington, D.C, in the summer of 2002,” wrote University of Toronto exoplaneteer Ray Jayawardhana in his 2011 book Strange New Worlds, “transits were all the rage. Keith Horne from the University of St. Andrews counted two-dozen transit searches in the works, employing a variety of instruments ranging from wide-field cameras that used commercially available 200-millimeter Canon lenses to the majestic 4-meter telescope on Cerro Tololo, Chile.” But at the time of that 2002 conference, he continued, almost three years after HD 209458 b had been confirmed, “no other discoveries had been reported.”
That wouldn’t change for another three years. Finally, in 2005, Greg Henry, under the N2K banner, picked up the transit of a planet called HD 149026 b, midway in size between Neptune and Jupiter. Just as in the case of HD 209458 b, the first transiting planet ever found, this one had originally been discovered by the radial-velocity technique, the planet-induced wobbling of its star, so that when Henry got the transit, the size and mass could be combined to figure out the planet’s density. Surprisingly—or maybe unsurprisingly, since the unexpected is routine in the world of planet hunters—the new planet was denser than anyone had anticipated. HD 209458 b had been less dense than anyone had expected. HD 149026 b was bigger than Jupiter, but was structurally more like Neptune, with a relatively large, solid core of maybe seventy Earth masses surrounded by a relatively smaller atmosphere.
Astronomers were finally beginning to get a handle on what exoplanets were actually made of, and with all these surveys and more in operation, they’d keep adding to the list of planets whose densities they could calculate. It was slow going, however. They had to look through Earth’s murky atmosphere, and our planet’s rotation kept them from being able to stare continuously at any one star. It would be impossible, as Batalha had said, to find a planet as small as the Earth orbiting a Sun-like star.
It would be impossible, at least, with either transit searches or radial-velocity searches. There was another way to look for planets, however, based on a theoretical idea invented, though never proposed as the basis for actual observation, by Albert Einstein. When Einstein came up with his theory of general relativity back in 1916, one of its implications was that space-time would literally be warped by massive objects. The Sun, for example, should cause enough of a distortion in space that a ray of light passing by would be forced to change course. If the Sun and a random star happened to be close together in the sky, light from the star, speeding straight as an arrow for most of its journey across the universe, would change direction slightly. The effect for observers on Earth would be that the star seemed slightly out of position, compared to where it would be if the Sun weren’t there. In 1919, astronomers took advantage of a total solar eclipse to try to measure the effect (there would be no point in trying at any other time, since you could never see nearby stars if the Sun were blazing away normally).
Sure enough, Einstein was right. LIGHTS ALL ASKEW IN THE HEAVENS, blared the headline in the New York Times on November 10, 1919. “Men of Science More or Less Agog Over Results of Eclipse Observations. EINSTEIN THEORY TRIUMPHS. Stars Not Where They Seemed or Were Calculated to Be, but Nobody Need Worry.”
Decades, later, when he had moved to the Institute for Advanced Study, Einstein received a visit from a Czech electrical engineer named Rudi Mandl. Mandl had evidently become obsessed with the idea that under the right conditions, this light-bending effect could turn a star into a lens, magnifying the image of a more distant background star. He raised some money, traveled to New Jersey, and knocked on Einstein’s door to beg the great man to confirm or deny this idea. Mandl was evidently relentless, possibly to the point of being somewhat annoying. So Einstein finally agreed, and it turned out that Mandl was right. Einstein wrote a short paper that appeared in Science in December 1936 that began: “Some time ago, R. W. Mandl paid me a visit and asked me to publish the results of a little calculation, which I had made at his request. This note complies with his wish.” Einstein also wrote a private note to the editor of Science, thanking him “for your cooperation with the little publication, which Mister Mandl squeezed out of me. It is of little value, but makes the poor guy happy.” In the paper itself, Einstein wrote that “there is no great chance of observing this phenomenon.”
As it turned out, Einstein was wrong. The phenomenon now known as gravitational lensing was first observed in 1979. Since then, this cosmic optical illusion has become one of the most powerful tools in astrophysics. It’s been used to measure the size of the universe, for example, and to study galaxies so faint they’d be invisible if they weren’t enhanced by the magnifying effect of closer galaxies.
In 1992, a Polish-born scientist named Bohdan Pacyznski came up with the idea of using gravitational lenses to figure out what dark matter, the stuff Vera Rubin and others had detected but not identified, actually is. (Pacyznski, who died in 2007, also came up with the idea of HATNet—originally to look for variable stars, subsequently hijacked to look for transits. He was also one of the first to suspect that so-called gamma-ray bursts are not feeble, nearby pops of energy but fantastically powerful ones happening halfway across the universe.) One possibility was that the dark matter consisted of subatomic matter, in the form of “weakly interacting massive particles,” or WIMPs. Another was that it was made up of brown dwarfs, smaller than stars but bigger than planets. Theorists had begun to call them massive compact halo objects, or MACHOs (“halo” meaning they surrounded the Milky Way). Whether these names should be considered brilliantly clever or a bit too cute is up to the reader.
Pacyznski realized that if MACHOs existed, they would occasionally drift in front of more distant stars, briefly magnifying the stars’ light through gravitational lensing. Paczynski called the phenomenon “microlensing” and he organized a survey called the Optical Gravitational Lensing Experiment (or OGLE—astronomers can’t seem to resist), run primarily out of the University of Warsaw, to look for them. OGLE didn’t find many MACHOs, but once exoplanets began showing up in the mid-1990s, it became clear that the technique would work for exoplanetology as well. If a star drifted in front of a more distant star, the latter would seem to spike in brightness for a short time, just as Einstein had said. But if the nearby star had a planet in tow, the spike might be followed (or preceded, depending on how things were arranged), by a dimmer spike as the planet drifted by in its turn.
If the planet were too close in to the star, the flares would come almost right on top of one another, so they’d be too hard to separate. So microlensing is especially good at finding planets—even planets as small as Earth—that are relatively far out. Transit and radial-velocity searches, by contrast, are best at finding close-in planets. That makes microlensing the best way to find Earths in the habitable zones of their stars, with one important caveat. A planetary system that randomly drifts in front of a distant star is only going to do it once. It won’t come around again, so you can never do any follow-up observations. You’ll never be able to measure a planet’s physical size, or study its atmosphere. “It’s not good for what I love,” Dave
Charbonneau told me, “which is characterizing the worlds I discover.” But like Kepler, it could be an important statistical tool, telling astronomers how often what sorts of planets occur in what sorts of arrangements.
Only about ten planets have been found so far by micro-lensing, but that doesn’t reflect poorly on its importance, says Charbonneau. “It’s not fair to compare it to the huge efforts in radial velocities and the huge efforts in transits,” he told me. “If we go ahead with a mission where there’s a significant microlensing element, it really could be transformative, because that would deliver thousands of planets that orbit far from their stars.” Beyond that, he added, since the foreground star and planet move across your field of view, you’re getting a snapshot of the system rather than waiting until the planet completes several orbits. “It’s great if you’re impatient, because it is the one method where you don’t have to wait for even one orbital period to know if you’re seeing a planet.”
Finally, in the late 2000s, Charbonneau himself began focusing his attention on a type of star that’s not at all like the Sun. Astronomers group stars into categories based on their surface temperature and their intrinsic brightness (that is, how bright they’d seem if all stars were equally distant from Earth). The Sun is a G-type star, with a surface temperature of about 6,000° Celsius. For every star like ours, however, the Milky Way has about one hundred smaller, redder stars known as M-dwarfs, which are only around half as hot. M-dwarfs are so dim that they’re very difficult to spot (astronomers haven’t even found all of the M-dwarfs in our own nearby cosmic neighborhood). They’re so much cooler than the Sun that their habitable zones—the orbits where water can remain liquid, and, in principle, nourish life—are much closer in. An Earth that orbited an M-dwarf once every 365 days would be frozen solid. An Earth orbiting close enough that its year was just a few days long would potentially be hospitable.
It turns out that both of these characteristics make M-dwarfs ideal places to search for small worlds. If you like looking for transits, the shadow of an Earth-size planet moving in front of a Sun-like star cuts out about 1/10,000 of the star’s light. If an Earth transits an M-dwarf, it cuts out 1/1,000 of the light—a much bigger signal, and much easier to spot. If you’re doing radial-velocity measurements, the fact that your Mirror Earth is so close, and also so much bigger compared with its star, means that it will yank the star around much more powerfully than we yank the Sun. And with both techniques, the planet’s shorter year means you don’t have to watch so long to see the orbit repeat itself. You can convince yourself it’s really a planet much more quickly.
With all of these advantages in mind, Charbonneau put together the MEarth (pronounced mirth) Project, which would look for planets just a bit bigger than Earth, around two thousand M-dwarfs within the closest few tens of light-years. “We think,” said Charbonneau when I asked him about it, “we have the sensitivity to get down to planets twice the size of Earth.” So he wasn’t quite looking for Earths; he was looking for super-Earths—a category of planet that nobody had thought much about. In our solar system, the next biggest planet up from Earth is Neptune, about four times as big across, and about seventeen times as massive, as Earth. A planet only twice as big as Earth might be similar in composition to Neptune, with a core of rock and ice surrounded by gases, or it might be mostly rock. If it’s the former, it would be hard to imagine life on such a world; if it’s the latter, maybe not so hard.
Or it might be something else entirely. From the moment the very first exoplanet was discovered, in 1992, astronomers have been ambushed over and over by the universe. The planets they find are not the planets anyone anticipated. The careful reader will assume “1992” is a typo that managed to sneak into print, since Michel Mayor’s first universally acknowledged exoplanet, 51 Pegasi b, was discovered in 1995.
What happened in 1992, however, is hard to classify. Everyone agrees that Penn State astronomer Alex Wolszczan and Dale Frail, of the National Radio Astronomy Observatory, found two Earth-mass planets in that year. Yet the discovery was so odd, and fits so poorly into any sensible narrative about the search for a Mirror Earth, that you inevitably have to add a but. It’s something like the old home run record in baseball. In 1927, Babe Ruth hit sixty homers, setting a record that went unbroken for decades. Then, in 1961, Roger Maris hit sixty-one. But Ruth hit his sixty in a season that lasted 154 games. By 1961, the season was 162 games long, and Maris didn’t hit the last one until the final game of the season. So did he really break the record? This is the sort of thing baseball fans can argue about endlessly.
In the case of Wolszczan and Frail’s discovery, the awkward part has to do with the star they were looking at. It’s a pulsar, the collapsed, super-dense nugget left over after a star has died in the titanic explosion known as a supernova—an explosion so bright that for a few days it can outshine the rest of the stars in the galaxy combined. For a really big star, the leftover is sometimes a black hole. For something smaller, it’s a chunk of matter just a few miles across, but more massive than the Sun. Just a teaspoonful of neutron star would weigh something like ten million tons. Some neutron stars rotate hundreds of times per second, and of these, some send out bursts of radio energy as they spin. When these incredibly rapid, precise blips were first picked up by radio astronomers in 1968, nobody had a clue what they were. They were briefly nicknamed LGMs, for Little Green Men, since at first nobody could think of a natural process that could cause such a precise, rapidly repeating signal.
What Wolszczan and Frail noticed was that the blips from a pulsar called PSR 1257+12 would vary slightly, coming closer and closer together, then farther apart, then closer. The best explanation was that something was pulling the pulsar toward, then away from the Earth. As the pulsar moves toward us, each blip of radio energy has just a little less distance to travel than the one before, so they come closer together in time. As it moves away, the blips have to travel a bit farther each time. The timing suggested two planets, each more or less the size of the Earth. It was crazy enough that Wolszczan and Frail held off on announcing the discovery, mindful of Richard Feynman’s dictum that “you must not fool yourself—and you are the easiest person to fool.” But while they were reanalyzing their data, a colleague named Andrew Lyne announced a planet around a different pulsar. Lyne’s discovery was published in the prestigious journal Nature with great fanfare. Wolszczan and Frail figured they’d blown it. Maybe they’d been too cautious.
As it turned out, they hadn’t been. A few months after his astonishing discovery had made headlines around the world, Lyne was preparing a triumphal talk about it for the upcoming meeting of the American Astronomical Society. Working late one night, he realized to his horror that he’d left out a crucial step in analyzing his observations. He knew, instantly and instinctively, what would happen when he redid the analysis. Sure enough, he told me shortly afterward, “the planet disappeared.”
It would have been bad for Lyne’s reputation if someone else had discovered the error before he did. It would have been even worse if he had insisted he was right after everyone else realized he was wrong. But Lyne found the mistake himself, and gave his talk as scheduled, but with a very different theme than he’d planned. The audience was utterly silent as he explained his mistake—and then, when Lyne finished, hundreds of astronomers gave him a standing ovation that lasted more than a minute. At the time, John Bahcall, Sara Seager’s old adviser, was the society’s president. After the talk, Bahcall came up to me and said, “I want you to know that Andrew Lyne’s talk was the most honorable thing I’ve ever seen. A good scientist is ruthlessly honest with him-or herself, and that’s what you’ve just witnessed.”
Lyne’s pulsar planet wasn’t real, but Wolszczan and Frail’s, it turned out, were. If the search for a Mirror Earth is ultimately a search for life, this discovery doesn’t fit anywhere. Life couldn’t possibly exist here. Planets couldn’t survive a supernova explosion, so they must have formed afterward, out of some of
the debris. But the neighborhood of a tiny, radiation-spitting neutron star would be about the most hostile environment possible. So the pulsar worlds are at once the first planets ever found and off at a tangent that makes them irrelevant to the search for a Mirror Earth. Exoplaneteers invariably mention the pulsar worlds when they talk about the history of exoplanet science, but don’t talk about them much at all when discussing the current state of the research.
The pulsar planets were pretty much completely unexpected. So were the hot Jupiters Marcy and Mayor began to find in 1995. At that time, the theory of planetary formation was based on the assumption that our own solar system was probably typical. It wasn’t obvious why this should be true, but it was even less obvious why it shouldn’t be. The idea was that a cloud of interstellar gas and dust collapsed, less than five billion years ago, spinning faster and faster as it got smaller and flattening out into a disk. The dense core of the disk formed into the Sun, whose heat drove the lighter material, including hydrogen, helium, and water vapor, outward toward the edges and left only the heavier rocky material closer in. The rocky stuff formed into the Earth, Venus, Mars, and Mercury. Farther out, the rocky material congealed as well, but there was more of it. If you mentally divide the disk into bands, the bands farther out are much bigger around, so they have more total stuff—more rocky material, plus the extra gas pushed outward by the newborn Sun. The rocky matter would have congealed into big, rocky planets whose gravity then started vacuuming up the lighter gases. The result is worlds like Jupiter and Saturn, with solid cores shrouded in massive, thick atmospheres.
A Jupiter or a Saturn couldn’t form close in, however, because there wouldn’t be enough rock to form its massive core, and there wouldn’t be enough gas to make the atmosphere. When 51 Peg b showed up, it was something like the situation when an elementary particle called the muon was discovered in 1936. No theorist had predicted such a particle, and the Columbia physicist Isidor Isaac Rabi responded by asking, rhetorically, “Who ordered that?” as though the muon were an exotic dish delivered unexpectedly from a Chinese restaurant. At least one theorist, however—Doug Lin, of the University of California, Santa Cruz—had predicted hot Jupiters. He’d argued, even before the discovery of 51 Peg, that in some cases a Jupiter-size planet could spiral inward, pushed toward its star by the gravitational effects of gas remaining in the disk. But most of his colleagues either didn’t know about it or thought it was just a clever theoretical exercise. Suddenly, in the fall of 1995, it was a plausible explanation for a discovery that had come completely out of left field.
Mirror Earth Page 13