Thursday, August 06, 2009

Kepler mission news: it works. It works really well.

The Kepler Mission, NASA's satellite dedicated to finding Earth-like planets, has just released its first science results via a news conference broadcast on NASA TV. Frankly, I was hoping to hear about planets that Kepler discovered during its early-mission check out, but the news conference did not focus on that. (I might know some things about those early results, but even if I do, I would not write anything more than what the press conference covered because I'd be getting some nice folks into deep trouble.)

Kepler works by measuring the brightness of stars. If a star has a planet, and if the planet's orbit takes it in front of the star as seen from Earth (only a very small fraction of planets will have that kind of orbit; it's all due to luck), then when the planet goes in front of the star (called a transit, it will block a tiny portion of the star's light. Planets the size of Jupiter block about 1% of the light of their parent stars, while Earth-sized planets only block about 0.01% of the light of the parent star. Since Kepler's goal is to find Earth-like planets, it has to demonstrate an accuracy of 0.01% (or 1% of 1%) for each of a hundred thousand stars. Today's news conference showed that Kepler is indeed getting data that accurate.

Here is a plot showing the brightness of a star seen by Kepler called HAT-P-7. HAT-P-7 is known to have a planet around it; the planet is called HAT-P-7b. HAT-P-7b is 1.8 times the mass of Jupiter and orbits its parent star every 2.2 days at a distance of 5.6 million kilometers (3.5 million miles), or less than 1/25th the distance between the Earth and the sun. The top line of the plot shows what Earth-bound telescopes see when they look at the star. A star with constant brightness should be just a straight line; instead you see a fuzz of points that is relatively straight. The fuzz is due to errors induced by Earth's atmosphere. But, in spite of the fuzz, you can see that the star is pretty constant, with a small but definite dip when the planet transits the star.

The second (bottom) curve is the same star as seen from Kepler. THAT is how much better Kepler is than telescopes on the Earth.

This plot shows the same Kepler data magnified 7 times (top) and 100 times (bottom). Magnifying it 7 times, you can just barely begin to see that the points are not a straight line, but vary a little bit. With the 100 times magnification, you can finally see Kepler's "fuzz." But you can also see a few other neat things.

First, you see a second dip in the light curve. The giant one is when the planet goes in front of the star. The little one is when the star eclipses the planet, also called the secondary eclipse. The fact that we can see this means that we are seeing light from the planet. It's just a tiny bit of light, but we can see it. And so when the planet goes behind the star, its light is blocked from view.

The other thing you can see in the 100-times magnified plot is that the light outside of eclipses is not constant, but it goes up and down as the planet orbits. This means that we are seeing phases of the planet, as it goes from "new" HAT-P-7b (the big transit) to crescent phases, to half phases, to a full phase (just before the secondary eclipse), and back to crescent and new phase again. Here's a movie illustrating the transit, secondary eclipse, and the phases.

Even more analysis of the amount by which the light is changing shows that the planet is not merely reflecting star light, like the planet Venus does as it goes through phases, but that HAT-P-7b is actually glowing. This isn't unexpected; at only a few million kilometers away from the surface of its parent star, the planet must be hellishly hot -- over 4400 degrees Fahrenheit (2300 Kelvin). At that temperature, the planet actually glows, just like the heating element on an electric stove.

In addition, when you look closely at the data, we can see that there is other stuff going on, too. Wiggles in the light are due to slight changes in the star itself. Our own sun does this, varying ever so slightly in brightness on time scales of minutes and hours. On the sun, these variations allow us to study the interior of the sun, so we should even be able to study the parent stars of every planet we find.

Lastly, this is just the first couple of weeks of data on this star. Imagine what we can do when we have three years of data!

So, when will we know if Kepler has found any Earth-like planets? Finding Earth-sized planets in Earth-like orbits takes time -- at least 3 years. Why? Because, in order to confirm a planet, mission scientists want to see the transit of the planet at least three times. The first transit tells you there may be a planet, the second transit confirms gives you an estimated length of the orbit, and the third transit confirms that the planet's orbit repeats with that period. This is crucial: suppose there are two Earth-sized planets around a single parent star at the distance of Jupiter. Each individual planet will only transit once every decade, but if there are two planets, we may get unlucky and see one transit and then the other a several months later. When a third transit doesn't happen months later, we'll know that we were faked out by something else. So, we need three transits, meaning three orbits, and the Earth takes one year to orbit the sun. So, that's three years. Stay tuned!

No comments:

Post a Comment