Mirror Earth (11 page)

But as early as 1992, he and his frequent collaborator Ron Gilliland, of the Space Telescope Science Institute, had already begun thinking about looking for planets by means of transits. It was mostly theoretical at first, since everyone still thought that other solar systems would resemble ours. If that was true, then the biggest, most easily detectable planets (like Jupiter) circle their stars in long, loping orbits. If they did transit, it would be only once every decade or more. Even then, the odds were long that they could spot such a transit: At a distance of hundreds of millions of miles from the star, the plane of a planet's orbit would have to be so precisely aligned with our line of sight that such a thing would happen only rarely. A huge planet orbiting right up against its star wouldn't have to be aligned with nearly so much precision—but such a planet was beyond the imagination of most astronomers at the time. “But then in '95,” said Brown, “when Mayor found 51 Peg in such a tight orbit, it became obvious that it was sensible to go looking. It was obvious to me, anyway. I had trouble convincing others.”

He and Gilliland had already convinced themselves, however, that they could achieve the necessary precision to detect transits, even without a huge telescope. Brown and Gilliland got together with Ted Dunham at Lowell Observatory—the same place where Percival Lowell had found “proof” that Mars was inhabited—to cobble together their first transit-search
telescope. (Dunham would ultimately become the Science Team director for the Kepler Mission before leaving to work on a high-altitude infrared observatory.) By 1999, they had it built, and Brown set it up in a parking lot at NCAR to run it through some tests.

It was just at this moment when Bob Noyes, Brown's old collaborator, was trying to help Dave Charbonneau find a thesis topic. Noyes knew about the transit-search project (Brown called it STARE, for STellar Astrophysics and Research on Exoplanets), and he suggested that Charbonneau head west to help out. “I bought a car,” recalled Charbonneau, “and I drove out to Boulder, and started working with Tim.” Before he went, he stopped to see Dave Latham. As Latham remembers it, “Dave ambushed me outside the door to my office as I was leaving one night and said, ‘I'm leaving for Boulder, and I need your advice on some good objects to look at.' ”

Although Latham had let Michel Mayor and Tsevi Mazeh pick up the planet-hunting project and run with it a full decade earlier, he was still in touch with them. Latham knew, although it hadn't been announced publicly, that Mayor and Mazeh had found a radial-velocity signal in a star called HD 209458. The size of the planet, and the nature of its orbit—assuming it was a planet, naturally—made it a prime candidate to transit across the face of the star. “Look at HD 209458,” Latham advised. “I'll tell you exactly when to look.” This was in late August. Charbonneau forwarded Latham's heads-up to Colorado, where, sure enough, Brown found the first transit. “I think it was 9/9/99, September 9, 1999,” said Charbonneau, as we walked from his class to his office. “That was the first
recorded transit of an exoplanet in front of its star.” Charbonneau showed up in Boulder a couple of days later, in time for the second transit, on September 16.

But Charbonneau and Brown figured this out only in retrospect, a couple of months later. That's because Brown had another search for transits going on at the same time, with his old partner Ron Gilliland. This one used the Hubble Space Telescope, in a sort of preview of the Kepler Mission. As Bill Borucki had realized ten years earlier, the chance of seeing a transit on a single random star (that is, one where Dave Latham hasn't tipped you off first) is vanishingly small. The best way to search is to look at a field of stars all at once. With its small field of view, Hubble isn't the ideal instrument to use, so Gilliland and Brown cheated a little: They focused on one of the hundreds of globular clusters that dot the central regions of the Milky Way.

Globular clusters are knots of up to a million stars—almost like miniature, spherical galaxies. The astronomers pointed the Hubble at a cluster called 47 Tucanae, and took data on about thirty-five thousand stars. 47 Tuc was much too far away for radial-velocity measurements, so there was no hope of getting densities for any planets they might find—but getting a sense of how many planets there might be in this big sample, and what sizes and orbits they came in, would still be useful. They predicted they might find seventeen transits. In the end, they found zero, perhaps because globular clusters contain mostly very old stars that have relatively little of the elements planets are made of.

Brown and Gilliland had gotten what Brown calls “a big pile of Space Telescope data” just as he and Charbonneau began doing their ground-based observations from the back parking lot. So they put aside the STARE data and began working their way through the Hubble data. Before they could get very far, Charbonneau got a call from back home. It was John Huchra, the director of graduate studies and former head of the Harvard astronomy department. Huchra was an unrepentant cosmologist—none of this planet stuff for him!—who had helped create the first large-scale maps of the universe in the eighties.

He was calling because Charbonneau had never gotten official permission to go west. Bob Noyes knew about the trip, but that wasn't good enough. “Huchra said, ‘Dave, you've got to come back and you have to make a case for the science you're doing. We're really nervous about students going away and working with scientists outside of the CfA because we can't supervise you.' I think maybe he thought I'd gone out to Colorado to go skiing, I don't know,” Charbonneau told me.

In any case, he rushed back to Cambridge. “I remember I went through a long defense—it was a couple of hours, at least.” About ten minutes of that was talking about why transits would be interesting; that part was pretty obvious. “The rest,” he said, “was what my thesis would say when I didn't find a transiting planet, which was what they all assumed would happen.” So he spent the rest of the time talking about the things they believed he
would
find, like binary stars and pulsating stars, and all the good science he could wring out of
those. “Of course,” Charbonneau continued, “the irony is that I already had the data that would show the transits of 209458, even though I didn't know it yet.”

The thesis committee was satisfied and Charbonneau headed back to Boulder. In the end, it took two months of working on the Hubble data before they could get back to their parking lot project. At almost exactly the same time, though, Brown got wind that Geoff Marcy and a collaborator named Greg Henry, at Tennessee State, had detected a transit as well, although he didn't know what star they were observing. “Geoff and I had an interesting phone conversation,” he recalled, “which amounted basically to ‘You show me yours and I'll show you mine.'” It turned out to be HD 209458. Brown and Charbonneau didn't get extra credit for the fact that their observations beat Marcy and Henry's by two months, since they hadn't analyzed them. The two teams published their discoveries in the same issue of the
Astrophysical Journal
the following year, and they're generally given equal credit for the discovery (although some websites mention just one or the other; a news release published on the UC Berkeley website at the time, for example, mentions only Marcy's group).

Credit aside, though, the observations marked another crucial step forward in the search for Earth-like planets. Four years earlier, Michel Mayor and Geoff Marcy and their teams had made a strong case that planets, or something that looked a lot like planets, orbited around Sun-like stars. With the discovery that at least one set of radial-velocity wobbles was matched by a series of transits, with precisely the same timing,
it became impossible to doubt that many of them, if not most, had to be planets.

Moreover, the fact that the planet now known as HD 209458 b already had a known mass, thanks to Mayor and Mazeh, and that its size was now known from the amount of light it blocked during its transits, allowed the astronomers to calculate its density. It was surprisingly low: HD 209458 b turned out to be about 70 percent as massive as Jupiter, but about 35 percent bigger. With such a low density, it was clearly made mostly of gases, but what kinds, and in what percentages, and with what implications for the makeup of smaller, life-friendly worlds, were still unknown.

Before we finished our conversation, Charbonneau wanted to tell me one more thing. “When you defend your thesis,” he said, “you have to have an outside reader, in addition to members of your department. I picked Bill Borucki.” At the time, very few of the astrophysicists at Harvard knew much about Borucki, a government scientist who didn't even have a Ph.D. But Charbonneau did. “I already knew that he was going to be PI of the Kepler Mission and I knew Kepler would find all of these planets, these Earth-like planets. Hopefully, anyway.”

When Dave Charbonneau arrived at Harvard, it turned out that he wasn't the first Canadian from the University of Toronto to show up there intending to be a cosmologist. Two years earlier, a woman named Sara Seager had come to Harvard with exactly the same intention. They'd known each other back in Toronto, but she had been two years ahead. “That doesn't seem like a lot now, but as you know, when you're younger it's a much bigger deal,” she told me one morning at her office at MIT. No one would ever describe the MIT campus as charming; it's stark and soulless compared with leafy, ivy-covered, Georgian-brick Harvard a couple miles to the northwest. But Seager's office is on an upper floor in the Green Building—a high-rise that is typically bleak on the outside, but with a view from the higher floors that plenty of Harvard professors would undoubtedly kill for. Her window looks down on the Charles River Basin, filled with sailboats and rowing shells for much of the spring, summer, and fall. Across the basin, the towers of downtown Boston loom only a mile or so away. Seager and Charbonneau had first met,
she told me, when she was a senior and Charbonneau was a sophomore. “He was thinking of dropping out of physics,” she said, “and after I graduated he sent me this really nice letter thanking me for convincing him not to.”

When she first got to Harvard, Seager worked on the recombination of the universe. This is the time, about four hundred thousand years after the Big Bang, when the hot, dense universe cooled off enough for atoms to form out of subatomic particles. The light that burst free as a result of that event is what astronomers stumbled on in 1965, proving the Big Bang had happened, and what cosmologists use to try to figure out how matter in the universe was arranged at the time. Like Charbonneau, Seager was working on the most popular topic in astrophysics in the mid-1990s—just as it was starting, like the early universe itself, to cool off. “My most highly cited papers are still the ones I wrote in cosmology,” she said. “But that's because there are more cosmologists to cite papers, just so you know.”

This is how a conversation with Sara Seager often goes. She's charming and gracious (which is a pretty generic quality, it turns out, among planet hunters), but she talks with a palpable intensity and focus. Her mind always seems to be in high gear, and she sometimes drops in an observation, like the one about cosmologists citing papers, that gives a satisfying flash of insight—but only for a millisecond, because by then she's already moved on to another point I don't want to miss. It happened again one time when she was talking about her father, who sparked her interest in astronomy when he took her to an amateur star party as a young girl.

“I remember looking at the Moon through a telescope,” she said. “It was unbelievable. Do you remember the first time you ever saw the Moon through a telescope? Do you remember seeing it recently? Or ever? Isn't it so incredible? It is just unbelievable.”

Anyway, she went on, her father was a doctor in a suburb of Toronto, and he got frustrated with the Canadian medical system, which kept exerting more and more control about what kinds of procedures you could do, and how much you could charge. So, she told me, he decided to leave medicine and go into the hair-transplant business. “He always wanted to have hair,” she told me, “because he lost all his hair by the time he was nineteen, and people told him he couldn't have it. You know,” she said, in one of her characteristic asides, “most men who have no hair don't mind. Only one percent of the men without hair wish they had it.” I have no idea whether this is actually true, but knowing Sara Seager I'm fairly certain she looked it up once, and never forgot it.

Seager's biggest contribution to cosmology is a detailed calculation of how the physics of recombination must have unfolded—a theoretical analysis that would help observers make sense of what the satellites were seeing. “So I finished that; it was done. So what next?” Her thesis adviser, Dimitar Sasselov, had been at the conference in Florence where Michel Mayor announced the discovery of 51 Pegasi b, and he'd been sufficiently excited that he had started to move from cosmology into exoplanet theory himself. “Of course, we were talking about that,” Sasselov told me during a visit to Cambridge many years later, “and Sara said, ‘You know, I like this cosmology stuff, but I really want to work on planets.' So we came up with a project for her.” At the time, Marcy and Mayor between them had found only a handful of planets, and three or four were “hot Jupiters” like 51 Peg b. The atmospheres of these planets were hot gases. The early universe was a hot gas. That meant Seager could apply some of her cosmology calculations to planetary atmospheres.