Mirror Earth (24 page)

This isn't the main reason Winn became an exoplaneteer, but as he went through grad school at MIT he felt the urge to do something a little more practical. He tried medical physics, but it didn't click with his personality, so he returned to lensing, and also did some work in condensed matter physics. After grad school, he did a stint as a science journalist, writing for the
Economist
for a year, and he considered abandoning science in favor of writing. In the end, he took a postdoc at Harvard, where he continued to work on cosmology. But just in the years since Seager and Charbonneau had departed, Harvard had become a hotbed of exoplanetology.

“There was all this excitement in the air about exoplanets,” said Winn. “The prospect was just emerging that we could study their atmospheres, and there were all of these other intriguing physics problems posed by multiple planet systems and close-in planets.” Back at Princeton, meanwhile, his old mentor, Ed Turner, along with many other senior astronomers, had turned into exoplaneteers as well. “It seems to me,” he said, “that the early days of lensing in the 1980s had that same feeling of excitement and newness as exoplanets. But exoplanets have more staying power because of the quest for life.”

So now, among many other duties, Winn was serving as the project scientist for a space telescope project called TESS, the Transiting Exoplanet Survey Satellite. “It would be a successor to Kepler,” he explained, “but looking across the whole sky rather than at one narrow area.” The trade-off would be that TESS would gaze at each of the two million stars on its list for a much shorter time—months, not years. It couldn't find a Mirror Earth, with an orbital period as long a year. The biggest thing in TESS's favor is that if JWST or even TPF finally goes into operation, it will have a nice, juicy set of planets to look at: The stars in the TESS catalog are much brighter, on average, than the ones Kepler is looking at. That makes radial-velocity follow-up with a telescope like the Keck much easier. It also makes it easier to study the light passing through or bouncing off the planets' atmospheres, so exoplaneteers can study their compositions. And ultimately, if future telescopes can image the planets directly, the fact that TESS planets are closer to Earth will make those observations easier as well. It is, said Winn, “the natural next thing to do.”

It's also far cheaper than a billion-dollar flagship mission. Kepler was a Discovery-class mission, limited to a budget of no more than $300 million. TESS was being proposed as a Small Explorer mission—a SMEX, in NASA's acronym-happy universe—which had to come in at under $200 million. The project's principal investigator, George Ricker, also at MIT, had first submitted a proposal to the agency in 2009, but it didn't make the cut. The team resubmitted at its next opportunity, however, in early 2011, along with twenty-two others;
the following September it was selected as one of five that would get $1 million for an eleven-month “concept study.” The best two of these would go on to launch, as early as 2016. If TESS loses out on the final round, the team might want to call Bill Borucki in as a motivational speaker.

As the Kepler team tried to remind reporters every time their satellite found a new planet, finding planets wasn't the goal—not individual planets, anyway. The goal was to determine how many stars, on average, have a Mirror Earth orbiting around them, a planet of about Earth's size, located in the star's habitable zone. If the percentage is high in the Kepler sample, that boosts the odds that there will be Mirror Earths close to us. If it's low, you need a flagship mission after all, which can look farther out into the Milky Way.

It all depends on a number exoplaneteers have begun, over the last few years, to call η
_Earth
(that's the Greek letter eta, so the term is pronounced either “eta Earth” or, more commonly, to make it clear that the word
Earth
is written as subscript, “eta-sub-Earth”). It's the fraction of Sun-like stars that have a Mirror Earth orbiting them—or that's one definition. “There are actually many different definitions,” Andrew Howard, a postdoc working with Geoff Marcy, told me during my Berkeley visit. “That one is just the narrowest.”

Sometimes, Howard explained, people use the term to mean
Earth-mass planets around Sun-like stars, without saying anything about the planets' sizes. Sometimes, as with Kepler, they mean Earth-size, whatever the mass. Sometimes the definition is expanded to include M-dwarfs, not just Sun-like stars. Ultimately, it doesn't matter all that much. The point is to figure out how hard it will someday be to focus in on a Mirror Earth and search for evidence of life. If eta-sub-Earth is around 10 percent, Jim Kasting says, and you're talking only about Sun-like stars, that means you should expect to find just three Mirror Earths within the nearest fifteen parsecs, or about fifty light-years—about three hundred trillion miles in all directions. “I'm actually optimistic,” he said, “that eta-sub-Earth is going to end up higher than 10 percent. I think it's going to be more like 20 percent to 40 percent, somewhere in there. But we will have to wait for the Kepler folks to tell us.”

He said this in the knowledge that Kepler was up and working and beaming down information faster than the team at Ames and their collaborators elsewhere could process it. None of this had been certain back in 2007, when NASA officials came to Geoff Marcy urging him to write a proposal to come up with a preliminary number for eta-sub-Earth from the ground. “The goal for that project,” Marcy said, “was always very clear.” He would use the Keck II telescope, which NASA had helped fund with the idea of finding planets, to survey 166 nearby Sun-like stars, looking for radial-velocity wobbles. “Same old same old,” he said, “except we would be looking at very high precision, and doing repeated measurements at very high cadence.”

“High cadence” means they took measurements frequently,
to delineate the curve of back-and-forth motion as accurately as possible. With Kepler, the cadence is
extremely
high—the satellite measures the brightness of all 150,000 stars in its field of view once every thirty minutes, and a small subset of 512 stars once a minute, the latter mostly to look for transit-timing variations. With this project, said Marcy, which he called the Eta-Sub-Earth Survey, the cadence would be one measurement, or sometimes two, every night, which is pretty fast for a survey that has to go from one star to the next to the next. “The stars were chosen blindly,” he said. “We selected them without knowledge about the planets that might or might not be around them.” Choosing stars where you know planets exist is a cheat. It makes the survey nonrandom. Even choosing stars you think are more likely to have planets would be a cheat—by picking stars high in metallicity, for example, which Debra Fischer had shown to be especially fertile planet-hunting territory.

“We knew we wouldn't be able to find planets exactly the mass of the Earth,” he said. “Our technique can't do that, even for the closest-in ones. You get very, very close, but you can't find planets that are Earth mass or smaller, and certainly not out at one AU.” An AU, or astronomical unit, is the distance Earth lies from the Sun, or about ninety-three million miles. A Mirror Earth around a Sun-like star has one Earth mass and orbits one AU out. In our solar system, Venus orbits at a little over .7 AU. Mars is at a hair more than 1.5 AU. Pluto, with a highly elliptical orbit, varies from just under 30 to nearly 50. “So the goal,” continued Marcy, “is to measure the fraction of stars that have very small planets in close-in orbits where our technique is very strong.” There were a handful of others
doing similar projects, he said. “The Swiss team is doing the best. They're doing very good work, and they've found more than we have.”

By the fall of 2010 when we spoke, the project had been under way for more than three years, and Andrew Howard had just written up the results to date in a paper that was about to come out in
Science
. “When all is said and done,” said Marcy, “cutting right to the bottom line, we surveyed the planet inventory from those larger than Jupiter all the way down to the smallest we could detect, which was three Earth masses, and found that there's an ever-increasing number of planets toward lower and lower masses, down to the smallest. The funny thing about this result is that for me, this is like a lifelong dream. It was just fifteen years ago that finding a Jupiter, any old Jupiter, was amazing. Here we have the distribution of planets down to three Earth masses. It's completely unbelievable that we have come this far.”

There was just one thing that worried him. “As exciting as this discovery and the new paper is, the theory of planet formation, albeit still adolescent at best, makes a distinction between the formation of rocky planets like Earth on the one hand, and the formation of Neptune and Uranus, which are mostly not made of rock, on the other.” Neptune and Uranus are roughly 50 percent water; the rest is gas. “They are literally water planets,” said Marcy, “with a rocky core of five to ten Earth masses.” In our solar system, there's a huge gap between Neptune and Uranus, at seventeen and fifteen Earth masses (which simply means seventeen and fifteen times the mass of the Earth), and Earth, at one. Somewhere in that gap,
there would presumably be a transition. Above a certain size, you form a water world. Below, you form a rocky planet. Nobody currently knows where that transition lies, but with Kepler, and more crudely, with the Eta-Sub-Earth Survey, you can start to fill in the gap and see. “We already have masses and radii from lots of planets,” said Marcy, “and I can tell you that a whole lot of them down to just a few Earth masses are fluffy. They are low density. They are reminiscent of Uranus and Neptune.”

“So this then goes back to our wonderful Howard et al paper,” he continued, “which on the one hand is a dream come true but on the other hand offers pause for some sobering thoughts.” The good news, he said, is that the number of planets is rising inexorably higher and higher as you go toward less massive planets, all the way down to three Earth masses. This is a good sign that with just a little more work the exoplaneteers will find lots and lots of planets with the mass of Earth. Presumably, said Marcy, “the Earths would have a rocky surface with continents and plate tectonics and maybe oceans and lakes—the vision that comes out of science fiction movies.”

But the final verdict, he continued, was not yet in. “I'm a little bit worried,” he said, “and from the Kepler data, I have reason to worry. From the theoretical side and from the observational side, we haven't answered the question yet about whether Earth-like planets are common.” Planets the size of Earth, he was saying, might well not be Earth-like in any other way. He also didn't have any strong evidence to support this worry, but he said, “I have some weak evidence.”

This evidence came from the theorists who try to create
virtual solar systems with computer simulations. The planets with lots of gas—Jupiter, Saturn, Uranus, Neptune—form quickly, in the first few million years, because after that the gas left over from the original collapsing interstellar cloud that formed the solar system disperses. “What's left over,” said Marcy, “is dust, dust particles, maybe some pebbles of up to a half inch in size, and according to theory it takes another hundred million years roughly, maybe fifty million years, nobody knows for sure, to form the Earths.” But those Earths turn out to be hard to form. The theorists let the leftover material whirl around and collide and stick together in their simulations, and, said Marcy, “you end up with Marses with no trouble, but you don't make Earths very easily.” Earth, he pointed out, is nearly ten times more massive than Mars, and it's hard to gather enough material together to create one—in the simulations, anyway.

“You can't bet your house on what the theorists tell you,” he said. He knew this from the experience of finding hot Jupiters, which few theorists had ever imagined. “But if this scenario is correct,” he said, “seeing small planets from the Uranus and Neptune category offers virtually no information about the rocky planets. They are theoretically completely different scenarios, utterly different.”

In fact, the question of how planets form, and how they should be classified, has been around since the first exoplanets were discovered. Some of the worlds Mayor and Marcy found in the first few years were six or more times as big as Jupiter. Above a certain threshold size, an object would presumably not be a planet anymore, but rather a brown dwarf—but what was the size? Some theorists suggested that brown dwarfs,
which form, like stars, directly from the collapse of interstellar clouds, could actually be smaller than some planets, which grow from smaller bits of gas and dust. And then there were the Pluto wars, which ultimately forced the smallest planet to be reclassified as a dwarf planet. Part of the argument there was that Pluto was more like a huge comet than anything else.

To this day, there's no really good definition of the term
planet
, although Marcy's colleague and Natalie Batalha's mentor, Gibor Basri, did his best to come up with one in 2003. In Basri's scheme, any solar system could be divided into three kinds of objects. A
fusor
was defined as an object that achieves nuclear fusion in its core during its lifetime (this would include not only all normal stars but also the pulsars where Alex Wolszczan and Dale Frail had found the very first Earth-mass planets, since pulsars were ordinary stars before they blew up). A
planemo
was defined as a round nonfusor. Pluto, its sister world Eris, and the asteroids Ceres and Vesta, among others, are planemos. So are the Moon and some moons orbiting the outer planets (including Jupiter's Enceladus, Io, and Europa; Saturn's Titan; Neptune's Triton; and more). But these moons wouldn't classify as planets because in Basri's scheme a planet was defined as “a planemo orbiting a fusor.” A planemo orbiting a planemo didn't count. Pluto, Eris, and the asteroids, on the other hand, which orbit the Sun directly, did. You've undoubtedly never heard of planemos or fusors, however, since the terms, for obvious reasons, never caught on.