NASA's Kepler spacecraft is starting to put the pieces together in its search for virtual Earth twins in other planetary systems. Kepler, which launched in 2009, is on the lookout for planets that are about the size of Earth and have temperate surface conditions. One half of that formula was realized on December 5 when mission scientists announced the discovery of a planet in the so-called habitable zone, called Kepler 22 b, a few times larger than Earth. Now Kepler has located its first two Earth-size worlds, and although neither are plausibly hospitable to life, it seems only a matter of time before the mission scores its ultimate goal.

The two new worlds orbit a sunlike star 950 light-years away called Kepler 20. One has dimensions almost identical to our own planet; the other is just 87 percent Earth's diameter. The planets, which by convention have been assigned the names Kepler 20 f and Kepler 20 e, respectively, are the smallest exoplanets for which diameters are known. Francois Fressin and Guillermo Torres of the Harvard–Smithsonian Center for Astrophysics and their colleagues announced the discoveries in a paper published online December 20 in Nature. (Scientific American is part of Nature Publishing Group.)

"For the first time, we've crossed the threshold of finding Earth-size worlds," Torres says. "The next step is having an Earth-size planet in the habitable zone."

Tens of millions of kilometers from Earth, the Kepler spacecraft carries out a relatively simple task. It tracks the brightness of more than 150,000 stars, watching for them to dim repeatedly as a planet passes in front of a star's face. For a planet as small as Kepler 20 e, that dimming is incredibly subtle; every six days, when Kepler 20 e completes an orbit, the starlight dips to 99.992 percent of its regular flux for just a few hours.

By tracking how much of a star's light a planet blots out, astronomers can make a careful estimate of the body's diameter. For larger worlds, they can also make a complementary estimate of the planet's mass by using ground-based telescopes that measure how much of a wobble the planet's motion induces on its host star. Witnessing such a wobble also confirms the presence of a real planet, as opposed to some other astrophysical phenomenon that causes regular fluctuations in a star's brightness. (One planet-mimicking phenomenon is a binary star system behind one of Kepler's target stars; when one member of the binary eclipses the other it causes a temporary dimming that can be mistakenly ascribed to the foreground star.)

But such mass measurements are not currently possible for planets as diminutive as Kepler 20 e and Kepler 20 f. So Torres and his colleagues used a relatively new software technique called Blender, which calculates the likelihood that what Kepler sees is caused by a planet. They concocted a range of false-positive scenarios to determine how many could produce the observed signal. With help from NASA's Pleiades supercomputer, the researchers analyzed close to a billion different scenarios, Torres says. The analysis found that Kepler 20 e is 3,400 times more likely to be a planet than a false positive; the planet-to-fake-out ratio for Kepler 20 f is 1,370.

Those numbers make a solid case for both objects as genuine planets. But in the absence of a measurement of the planets' masses, their compositions remain unknown. Given the comparable dimensions of both newfound planets and Earth, a similar composition of silicate and iron is a possibility for either world, the researchers say.

4 Comments

1. 1. sault 02:12 PM 12/20/11

   2 thoughts:
   
   1. Is it possible to determine the orientation of a star's equatorial plane with current techniques? Does Kepler stare at all those stars and it just happens to get lucky when the planet's orbital plane intersects the Solar System? Or was the large sample size of stars analyzed beforehand to be made up of stellar systems whose orbital plane does intersect the Solar System?
   
   2. Is anybody trying to model what some of these systems would look like and how many other planets are possible in these systems. The configuration around Kepler 20 might rule out any smaller planets further out that wouldn't cross in front of their parent star as often. Or it might not. The orbital dynamics of the planets we do know might preclude the existence of planets at different orbits or it might hint at their presence.

   Reply | Report Abuse | Link to this

2. 2. bigbopper 03:09 PM 12/20/11

   Given the fact that there are hundreds of billions of stars in the average galaxy, and hundreds of billions of galaxies in the known universe, and that our investigations so far indicate that a large percentage of stars have planets, I agree that it's only a matter of time before we discover planets almost exactly like Earth.
   
   We're now generating reasonably reliable estimates for the first two factors in the famous Drake equation, namely the average rate of star formation per year in our galaxy, and the fraction of those stars that have planets. The next factor, the average number of planets that can potentially support life per star that has planets, should be within our grasp after the next big space telescope is launched which can spectroscopically analyze the atmospheres of individual extrasolar planets. Then there will be only 4 factors left to pin down. Should only take a few million more years!
   
   

   Reply | Report Abuse | Link to this

3. 3. jeffpc 08:48 PM 12/20/11

   I would imagine that it should be possible to calculate an expected time when Kepler will produce an earth like planet in the Goldilocks zone. Obviously large and orbitally frequent planets should be first because large create greater light variation and frequent create more measurements to remove uncertainty. Surely we can extrapolate how long before Kepler will remove sufficient uncertainty on an Earthsize world in a Goldilocks zone. My prediction is the first will be closer to its star than ours and that that star will be smaller. Anyone have sufficient information to make the calc?

   Reply | Report Abuse | Link to this

4. 4. Wayne Williamson 04:33 PM 12/21/11

   very cool(ok hot)...and as the other posters noticed...this only works on systems that happen to have their plane in line with us....

   Reply | Report Abuse | Link to this