Searching for Earth

NASA’s Kepler spacecraft will search for Earth-size planets around more than 100,000 stars.
Credit: NASA

In the past decade, astronomers have found more than 250 planets orbiting distant stars. Most of these have been giant planets, the size of Saturn or Jupiter, or larger. A few have been Neptune-size. No-one, however, has yet found the holy grail of extrasolar planets: an Earth-size planet with an Earth-like orbit around a sun-like star.

That’s about to change. NASA’s Kepler mission, selected in 2001 as the agency’s tenth Discovery mission, is a space-borne telescope designed specifically to look for alien Earths. It is “NASA’s first mission capable of detecting Earth-size and smaller planets around other stars,” says David Koch, an astrophysicist at NASA Ames Research Center near Mountain View, Calif., and the Deputy PI on the Kepler project.

Koch and his colleagues have successfully completed a critical test of the telescope’s imaging system, using hardware identical to what will be used on the spacecraft.

Kepler will look for planets that “transit” – move across – the faces of their stars as seen from our solar system. Kepler won’t actually see a transiting planet; it won’t take an image of it; rather it will measure the slight dimming of starlight caused by the planet moving across the face of the star.

The dimming caused by a transit is miniscule: one-hundredth of one percent for an Earth-size world. For comparison, imagine you’re inside looking out through a window. If you open the window and look directly out, the change in light intensity is about one percent. Kepler’s detectors will find Earth-like planets by measuring changes more than one hundred times as small.

Because this effect is so subtle, Kepler will be launched into space in an Earth-trailing orbit. “To look for transits, you have to get out into space. The reason is the atmosphere,” Koch says. When stars twinkle in the night sky, their brightness changes by much more than the change caused by a transit. “You can find giant planets using the transit method from the ground,” Koch says. But to find other Earths, the telescope has to be put into space.

Mercury can be seen transiting the sun in this image captured by NASA’s Solar Optical telescope onboard the Japanese satellite Hinode.
Credit: Hinode JAXA/NASA/PPARC

A planet’s orbit around its star can last for anywhere from a few hours to several years. Transits, however, last for only a few hours. For example, Koch says, “if you were to get back away from our solar system and look at the Earth transiting our sun, and it went right across the center of the disk of our sun, that would take 13 hours.” By noting the timing of a sequence of a planet’s transits and knowing the mass of the star that the planet orbits, the Kepler team will be able to calculate the planet’s distance from the star. How much a star’s light dims during a transit will indicate how large the planet is. Larger planets block more starlight.

Kepler will observe a group of more than 100,000 stars for four years, recording stellar brightnesses every half-hour. Kepler’s target star field, Koch says, is a region about the size of “two dips of the Big Dipper,” located between two of the brightest stars in the northern summer night sky, Deneb and Vega.

Because transits are such short-lived events, the spacecraft will watch the same patch of sky non-stop. “The whole idea for the mission was to look at one place in the sky continuously. We don’t look anywhere else,” Koch says. “If you blink, you’ll miss a transit.”

The first planets Kepler will find will be those closest to their stars. “Anything with [an orbital] period of a week or less, we will see four transits in that first month,” says Koch. “As time goes on,” he says, “we will be able to detect planets further and further out from the star. So after the first month’s worth of data we’ll detect planets that have periods of a week. After the first year, we’ll detect planets that have periods of a few months. And the reason for the four-year mission is so that, after four years, we will have detected four transits of a planet in the habitable zone, like Earth is in the habitable zone” of the sun.

The Kepler team will be looking for four evenly spaced transits before they feel confident they’ve found a candidate planet. But additional checks, using ground-based telescopes, will still be necessary to rule out possible alternative explanations.

Only about 5,000 stars in Kepler’s field of view closely match the sun’s characteristics and are bright enough to be able to detect planets exactly the size of Earth in the habitable zone. But for most of stars Kepler will monitor, the planets orbit in planes that are not aligned with the line of sight to the Earth; that means no transits. Kepler’s designers expect to detect about 50 planets as small as Earth in the habitable zones of their stars, assuming stars have both an Earth- and a Venus-size planet. They’ll detect hundreds if most stars have smaller planets close-in and if there is an abundance of super-Earth-size planets.

These discoveries would be a clear indication that planets like ours are common in our galaxy. But, says Koch, not finding those 50 planets will also be a meaningful result. “If we expect 50 and we get nothing, or 1 or 2, then we can say, you can know, Earth-like planets are not common. Just as profound a result.” In either case, once its four-year mission is complete, scientists will have a far more detailed picture of the distribution of planets in our galaxy than they do now.

Engineers assemble Kepler’s primary mirror.
Credit: Ball Aerospace

The test that Koch and his colleagues performed recently at NASA Ames was a final check of one of Kepler’s CCDs, along with its associated electronics. The test used flight hardware identical to what will fly on the spacecraft. Kepler will have 42 CCDs, each about 1 x 2 inches, containing a total of 95 megapixels. By comparison, the CCDs on digital cameras are about the size of a thumbnail and even top-of-the-line professional cameras typically contain about 10 to 12 megapixels.

To perform the test, Koch had built an artificial star field containing 1600 faint “stars.” Each star was actually a tiny rectangular hole, 10 microns wide, cut by a laser into a metal plate. Across some of these holes he mounted a wire about the width of a human hair. A light source was placed below the plate, and the CCD above it. The entire assembly was enclosed within a 10-foot-high vibration-resistant chamber surrounded by 4-inch-thick thermal insulation and thermal-electric heaters and coolers to maintain a constant temperature to within 50 thousandths of a degree.

To simulate a transit, Koch ran current through one of the wires, causing it to expand by a mere 12 nanometers. This blocked some of the light – a mere 80 parts per million – shining through the rectangular hole, just enough to look to the CCD like a transit will look to Kepler.

The entire test took several weeks to run. “A transit lasts a fraction of a day. So you can’t just take the measurement for five minutes,” said Koch. You have to “show that you can have this thing stable for many days. We ran tests as long as two weeks continuously” in some cases, to ensure that the setup was operating stably. “We didn’t want to have another Hubble,” Koch said.

The test was a resounding success. It showed that Kepler’s CCDs will work as anticipated. What remains now is to complete the spacecraft’s assembly and to launch it into space. The launch is scheduled for February 2009.


Related Web Sites

What Would ET See?
Astronomers Find Habitable Earth-like Planet
Finding Earth’s Twin
Searching for Pale Blue Dots
The Planetary Horizon