Our failure to hear transmissions, in other words, doesn’t mean they’re not out there. The men and women who have continued to listen for alien radio signals for the past fifty years and, more recently, for possible flashes of light from alien signaling lasers, have always been mindful of Philip Morrison and Giuseppe Cocconi’s observation that the probability of success is difficult to estimate. Even if millions of habitable planets exist, which is looking more likely all the time, and even if some of them have given rise to technologically advanced life, nobody really knows whether radio waves or laser beams are the standard way of talking across interstellar space. Maybe what we think of as advanced communications is their equivalent of smoke signals—a stepping stone on the way to technologies we can’t yet imagine.
Chapter 16
A WORLD MADE OF ROCK, AT LAST
By the time the American Astronomical Society’s winter meeting rolled around in January 2011, the astronomers who had gathered in Seattle were itching to get their hands on the four hundred planet candidates the Kepler team had dangled, then snatched away, six months earlier. It wasn’t quite time, though. Bill Borucki had promised the four hundred would be set free on February 1, and he wasn’t going to move the date forward.
The conference-goers wouldn’t have to leave empty-handed, however. In what was turning out to be a pattern, the Kepler team wouldn’t release candidates, but they would announce actual, confirmed planets, just as they had at the Washington meeting a year earlier. In Washington, the big news was simply that Kepler could find planets, and that radial-velocity measurements could figure out the masses and densities for some of them. The only announcement in the year since then had been Kepler-9, the first planetary system where the mass of the planets had been measured, not with radial velocities, but with transit timing, the powerful new technique nobody had even thought about when the Kepler Mission had first been approved.
This time, Natalie Batalha would be making the big reveal, at a press conference on the first day of the meeting. She agreed to brief me the day before, however, as long as I promised, on pain of her extreme disapproval, that I wouldn’t say anything about it before the official presentation. We met at a hotel a few blocks from the main conference hotel; the Kepler team was having a preconference workshop to talk about the upcoming data release the next month and to work through various routine issues that inevitably came up in analyzing and following up what the satellite was telling them. When I showed up, an Ames press officer went into the workshop to get Batalha. I could see a couple dozen astronomers, including Dave Latham, Dave Charbonneau, and Bill Borucki, listening to someone making a presentation. I asked if I could just sit in on the meeting. The press officer and Batalha both looked startled for a moment, then realized I couldn’t be serious. We all knew there was no way I was getting inside that room.
“So what we’re announcing tomorrow,” she said, when we sat down, “is the discovery of our first rocky planet, Kepler-10b. It’s in a very short period of less than a day, which makes it very similar to CoRoT-7b, if you’re familiar with that result.” What made this discovery special, though, she said, “is that our error bars on all of our measurements are very tight. We know unquestionably that this is a rocky world.” An error bar is another term for uncertainty, the plus-or-minus that goes with astronomical measurements of pretty much anything—the age of the universe, the distance between the Earth and the Sun, whatever. “And the reason that we know,” she said, “is because we’ve got all of our best capabilities coming together for this one discovery. We’ve got this exquisite photometry [the change in brightness as the planet moves in front of its star]. We’ve got a very high-precision radial velocity. And this star is bright enough that we can use astroseismology, which allows us to derive the fundamental stellar properties to within 2 to 6 percent accuracy.”
The team had known about this planet, or, at least, had known it was a good candidate, almost as soon as Kepler was launched. “We actually saw it during the commissioning phase,” she said, “before science operations had formally started.” Geoff Marcy’s group did its first radial-velocity measurements from the Keck in August 2009, about the time Marcy was willing to admit publicly only that the satellite was working properly. “So the mass of this planet is 4.65 Earth masses,” Batalha continued, “and the radius is 1.4 times that of the Earth.” The density, she said, works out to 8.8 grams per cubic centimeter, which makes Kepler-10b half again as dense as Earth. “This seems kind of high,” she said. “If you google it, you’ll find it’s about the density of an iron dumbbell.”
But that doesn’t mean it’s made of iron. Kepler-10b is so massive that it crushes itself down under its own weight. If you take the material the Earth is made out of—the same proportions of silicate rocks and iron and nickel—and just add more of them, the planet grows heavier at a faster rate than it grows larger, thanks to the crushing; and at 4.5 times the mass of the real Earth it would in fact be about as dense as a dumbbell. A dumbbell that heavy would be even denser. If you graph the size versus the mass of an Earth-type planet, and locate Kepler-10b on the graph, she said, “the error bar almost kisses that line of Earth composition. It’s just a little higher in density, so it does seem to be a little bit iron-rich. Kind of like Mercury.” CoRoT-7b kisses the Earth-composition line too, she said, echoing what Marcy had told me earlier. The difference is that, unlike Kepler-10b, it has huge error bars. “We’ve got three different papers on CoRoT-7b with mass estimates that range from, like, zero to ten. Zero is a funny way of saying it, but it’s basically a nondetection all the way up to ten Earth masses.”
For that reason, she was arguing, Kepler-10b, not CoRoT-7b, should be counted as the first unambiguously rocky world ever found beyond our solar system. It was too big and far too hot to be a true Mirror Earth, but it was another step closer. The star itself, Kepler-10, is about 560 light-years from Earth, which got Batalha thinking. “I told this story to the science team this morning,” she said. “Just out of curiosity, I subtracted 560 from the year 2010 and got 1450. The year 1450 is when the light left the star. And I googled the year 1450 to see what was happening—in Wikipedia it’s listed as being the beginning of the age of discovery. Isn’t that amazing? I love that. So Europeans were crossing the Atlantic for the first time when light left the star. I thought that was kind of a nice tie-in.” (It was now 2011, but since the distance to Kepler-10 isn’t known down to the light-year, it’s a legitimate bit of scientific poetry.)
A videographer named Dana Berry, who has done work for National Geographic, had been commissioned to create a video animation of what Kepler-10b might look like—a rocky planet, hugging its star with one face always turned to the searing light glowing red with heat. “When I saw it,” Batalha told me, “I thought about our small telescope up at the Lick Observatory.” This was the Vulcan telescope, which Bill Borucki used to prove to NASA that he could detect multiple transits, named for the planet astronomers once thought orbited Sunward of Mercury. “So when I saw the animation,” said Batalha, “the first thought that came to my mind was, ‘Wow, this is our planet Vulcan.’ ”
Whether it might ever be named Vulcan in a formal sense is unclear. So far, the International Astronomical Union, whose job it is to assign official names to heavenly objects as opposed to the numbers and letters different catalogs use to list them, has not come up with any sort of exoplanet-naming convention. According to Geoff Marcy, his wife had an idea back in the 1990s, when he started finding his first planets. “She said I should call them Susan 1, Susan 2, and so on,” he told me. Although it should be obvious that she was kidding, it’s probably a good idea to make it explicit: She was kidding. When 51 Peg b was discovered by Michel Mayor, a suggestion was floated that the planet be named Bellerophon, the name of the mythological Greek hero who tamed the flying horse Pegasus. One of the pulsar planets was unofficially named Methuselah (the planet was probably very old), and HD 209458 b, the first transiting planet, was called Osir
is, after the Egyptian god, but these, too, were never considered official. I’ve rarely if ever heard them used by astronomers.
When the press conference rolled around the next morning, one more thought had evidently come to Batalha’s mind. Or more likely, it wasn’t a thought, but something subconscious that led her to speak of the planet in an unusual way. “Today,” she said to a group of reporters, anxious for the latest Kepler news, “we’re announcing a new planet. She orbits her star …” Throughout her talk, Batalha referred to Kepler-10b as “she.” When it came time for questions, my hand shot up. “How,” I asked her, “did you determine the gender of the planet?” The room broke up and Batalha looked a little embarrassed. I was worried that I’d stepped over the line, even though I was obviously teasing, but Geoff Marcy, who had been up on the podium with her to provide commentary, assured me that I hadn’t. “That was a great question!” he said to me after the press conference was over. “I was surprised because she was so consistent. I was like, okay, I heard it once, now I just heard it again … I don’t know how much of a gender issue we still have in astronomy, since there are a lot of women at this meeting. But then,” he said, after thinking for a moment, “astronomy faculties are still mostly male, and the Kepler team is more male than female by a large margin. So the issue really hasn’t gone away, and I think the men and women of Natalie’s generation are still rightly sensitive to the issue.”
Kepler-10b was the closest thing to a Mirror Earth the Kepler team or anyone else had yet found, but it was only one of the four hundred planet candidates that had been held back the previous June. Just about three weeks after the Seattle meeting, the candidates were released at a press conference at NASA headquarters. When Bill Borucki began speaking before a crowd of reporters, NASA officials, and TV cameras, however, he didn’t just announce the four hundred; in the intervening half year, the team had vetted hundreds more. When the press conference was over, the number of planet candidates had leaped from 706 to more than 1,200. They broke down this way: 19 were larger than Jupiter; 165 were about the size of Jupiter; there were 662 Neptunes; 288 super-Earths; and, finally, 68 planets about the size of Earth, a few of which were a bit smaller than Earth itself. Slice it a different way, and you had 54 planets in their stars’ habitable zones, 5 of which were approximately Earth-size (all the members of this last group were orbiting M-dwarfs, whose habitable zones were very close to the stars and where you could see the required minimum of three transits in just a few months).
This was the piñata Dennis Overbye was talking about in the New York Times, a harvest of planets so overwhelming nobody knew what to do with them all. If Kepler had been looking across the entire sky rather than at a tiny patch, Borucki said, it probably would have found about four hundred thousand planets. That’s only in the nearby sky, not the Milky Way: Kepler is a powerful telescope, but not powerful enough to find planets much more than a few hundred light-years away. The Milky Way is about a thousand times bigger than that.
Kepler had found potential planets, Borucki emphasized; most of them would never be fully confirmed. Nevertheless, statistics based on the planets that had been confirmed suggested that about 90 percent of these objects were real—and these confirmed planets had come from only the first six months of Kepler observations. Having so many candidates brought the Kepler scientists a long way toward their goal of calculating the frequency of Mirror Earths. The team would continue to try confirming as many planets as it could, of course, with every trick it could come up with. Kepler’s primary goal was statistical, but finding real planets that could be studied, either with existing telescopes or with the more powerful telescopes that might be coming over the next decade, was a close second. The day after Bill Borucki’s press conference, in fact, the cover of Nature featured an artist’s rendering of the Kepler-11 system—the next one to be confirmed after Kepler-10b, Natalie Batalha’s hot, rocky Vulcan. This was the six-planet system for which Daniel Fabrycky had created a simulation that so fascinated his children—five planets ranging in size from super-Earth all the way up to near Neptune size, all huddled up against their star. If you took five planets bigger than Earth and crammed them inside the orbit of Mercury, you’d have Kepler-11.
“I was the one who first saw this system in the data and said, ‘This is really interesting, we should put a lot of effort into it,’” Jack Lissauer, the Kepler scientist who served as lead author on the Nature paper, told me the day before it was formally released. That effort included a campaign to monitor the star itself very carefully, to be absolutely sure about its physical size, and to make calculations, based on transit-timing variations, to confirm the planets and get their masses—the job of Fabrycky, the paper’s second author. (There were thirty-nine authors in all, including Eric Ford at number three; Bill Borucki, fourth; Geoff Marcy, sixth; Natalie Batalha, eleventh; Dave Charbonneau, sixteenth; Dave Latham, thirtieth; and Dimitar Sasselov, thirty-seventh. It was a who’s who of exoplanetology.)
What made the system so intriguing was partly the question of how it came to exist in the first place. All of the planets were presumably born farther out and migrated or were flung into their present positions. Packing them in so tightly, however, wasn’t easy for theoretical models to do. Another surprise was that the planets all orbited in precisely the same plane. “This was also going to be tough to explain,” said Lissauer. “Our solar system is approximately flat, but given the chaotic process that gives rise to planetary systems, it’s hard to see how this one came together. It’s so unexpected, and we’re getting so much information from this system,” he said. “I think this is the biggest thing in exoplanets since the discovery of 51 Peg b in 1995,” he continued. “My view is that a true Earth-mass, Earth-size planet in the habitable zone is the only thing that could be more interesting than this system. But until we find that, they’ll all just be hot rocks.”
So far, the Kepler team had been talking about two categories of exoplanets. There were candidates, more than 1,200 of them as of early 2011. And there were confirmed planets—about 20, distributed around seven stars, and confirmed by either a radial-velocity signal or evidence from transit-timing variations. In May 2011, at the smaller, spring meeting of the American Astronomical Society, they began talking about another category: validated planets. These weren’t quite confirmed, but had passed enough tests to rule out false positives that you could consider them planets anyway. The chances were just too small that they could be anything else.
One way to validate a Kepler planet was to observe a transit in infrared light, with the Spitzer Space Telescope—Dave Charbonneau’s project under the Participating Scientist Program. The dip in light from an actual transit should look the same with Spitzer as it did with Kepler; an eclipsing binary should show a color shift. Another newer and more rigorous technique, however, didn’t involve any observations at all. Instead, it relied on a computer-simulation technique called Blender, originally created by Harvard astrophysicist Guillermo Torres to rule out false positives in the OGLE survey. This was the ground-based survey originally designed to look for chunks of dark matter in the Milky Way, then redirected to look for the Einsteinian microlensing, or magnification, effect of distant solar systems on even more distant stars, then redirected again to look for transits.
As Bill Borucki had realized early on, a dip in starlight could be the signal of a planet, or it could be coming from a pair of eclipsing binary stars peeping over the shoulder of the target star—a blend of light from several stars, mimicking a transit. “OGLE,” Torres said at the meeting, which was in Boston, “looks at a very crowded field of stars, where the likelihood of a blend is much greater than with other surveys.” So he came up with Blender, which took all possible scenarios—a wide range of combinations of foreground and background stars of all brightnesses and sizes and relative distances from Earth—and used them to calculate how likely it was that the signal the telescope saw represented one of these combinations.
“W
hen Kepler came along,” Torres said, “I realized that they wouldn’t be able to confirm most of their candidates.” He approached the team and proposed they use his simulation. “They were very enthusiastic,” he said, “and it’s been very successful.” The first success was announced by his Harvard colleague Francois Fressin at the meeting. Using Blender, Fressin and Torres were able to validate the fact that Kepler-10b, Natalie Batalha’s hot, rocky Vulcan, has a companion, 10c, just a little bigger, at about 2.2 times the radius of Earth, with a forty-five-day orbit. With Blender, said Torres, “we’re fairly confident that what we’re looking at is a planet.” By fairly confident, he explained, he meant that “we’re looking for odds ratios of something like a thousand to one. That’s when we call it a validated planet. It’s not confirmed: There’s a difference in the terminology here.”
Torres and his colleagues also used Blender to validate the sixth planet in the tightly packed, CD-flat Kepler-11 system. “We’re now working on other candidates,” he said, “which have already been submitted for publication.” And they’ve used Blender to validate the existence of CoRoT-7b, the possibly rocky planet that was possibly found before Kepler-10b—although Torres, like everyone before him, couldn’t nail down its mass. Could Blender turn out to be the only way to validate a true Mirror Earth in Kepler’s list of candidates?
His straightforward answer: “Yes.”
Chapter 17
ASTRONOMERS IN PARADISE
If you ask me, it’s best to arrive at the Jackson Lake Lodge, about thirty miles north of Jackson, Wyoming, at night. The lodge is just inside Grand Teton National Park, so you know there must be some pretty good scenery out there somewhere, but in the dark it’s purely hypothetical—and unless there’s a full Moon out, it’s really, really dark. When you check in at the front desk, it feels like an ordinary hotel, aside from the western-themed artwork and the decorative wagon wheels and antlers sprinkled around the lobby. Most of the guest rooms aren’t in the main building, but rather in a series of cottages fanning out along small, poorly lighted roads. If you haven’t got a car, they hand you a flashlight so you won’t get lost. The bears, the clerks promise, are very unlikely to eat you. Most of the cottages are nestled in stands of pine trees, so even in full daylight, as you walk back to the lodge for breakfast the next morning, your sense of nature may be limited to the clean air and woodsy smell.
Mirror Earth Page 21