Friday, June 18, 2010
How many exoplanets!!!
As everyone knows, I love exoplanets, I marvel at the weird and wonderful worlds our telescopes have revealed, and quite often make Celestia files to share them (or at least their simulations), with the world.
As of 15 June there were 461 extrasolar planets, after CoRoT had revealed a swag of 6 new exoplanets. How was I going to find the time to make Celestia files for all of THEM I mused.
Then on June 16 the Kepler mission announced 706 exoplanet candidates, nearly tripling the number of exoplanets we know about. Thanks Kepler! Well, okay they have only released 312 candidates around 306 stars, with 5 systems with multiple planets, and held back 400 for further follow-up, but nearly doubling our known exoplanet roster in one go doesn't help me much.
Transit light curves observed by Kepler. Image Credit NASA.
Note the "candidates" though. Kepler detects the faint dips in light from a star that occurs when a planet goes in from of it. However, there are other things that can cause transit-like dips (sunspots, other stars in a close binary system etc.).
While the transits have been vetted for consistency, planet-likeness and for most, observed for at least 3 transits, they haven't been followed up by other measurements (eg radial velocity measurements to eliminate double stars systems and so on). These 306 systems that have been released are "low priority" for follow up, and will contain an unknown number of false positives. The 400 that have been held back are "high priority" and are being subject to further study as I type. Scheduling radial velocity and other measurements for 400 stars is no trivial matter, with telescope time being precious, and the 400 held over stars will be released in February 2011.
Figure 2 of Borucki et al., 2010, comparing the sizes of 306 planets found with Kepler against the sizes of planets found by other surveys.
Well, what of the the planets they have found. When we first found exoplanets, back in 1995, they were jaw-dropping hot super-Jupiters screaming around their suns in a handful of days.
This was pretty amazing, as planets that big shouldn't form that close to their parent stars. We soon found lots of other "super Jupiters", because if you are looking at how a star wobbles, or the dip in a stars light, it's easier to find something big orbiting often then it is to find something small orbiting slowly.
For example, if you were looking at the Sun for Jupiter, you would have to watch for 12 years to detect it, and you would need to wait 24 years to confirm your detection. So it is no wonder that our explanet detections up until now have been dominated by massive planets orbiting closer to their suns then Mercury is to ours.
The Kepler candidates change all that. If you look at figure 2 above, you can see that there are relatively few Jupiters (11 times Earth radius [Re] and above), lots of "Neptune-sized" planets (5-10 Re, with 58% of all planets being between 4-5 Earth masses) and a goodly number of Super-Earths (1.5 -4 Re with 12% of all planets being between 1.5-2 Earth masses). It looks as if close "Super Jupiters" are a rarity, and smaller worlds, more like Earth, are to be expected.
Figure 8 of Borucki et al., 2010, comparing the frequency of various planet types against star type.
The vast majority of these worlds have very short orbital periods, the median being 9.7 days (remember, Mercury takes 88 days to go around the Sun). There's an interesting "mini-solar system" around star catalog number 5972334, with a 2 earth-mass planet screaming around its G class sun in 2.4 days, and a Jupiter mass planet moving more sedately around in 15 days.
Of course, given the length of time Kepler looked at any one spot (33.4 days), finding planets with longer orbital periods is hard. You can infer by the length of the transit what the orbital period is, even if the period is longer than the observation time, and there is one Jupiter mass planet with a period of 206 days, similar to Venus, and another Jupiter mass planet listed as having an orbit around 28 years. As it orbits a star 9 times the size of our Sun that makes it rather Jupiter like indeed.
There is another factor making it hard to detect planets with longer, more solar system-like orbits. The further away the planet is from its Sun, the less likely it is for its orbital plane to be sufficiently closely aligned with Earth for a transit to be detected.
Image Credit, NASA.
Kepler is looking back along the Orion spur in an area between Cygnus and Lyra. It's deliberately sampling more "sun-like" stars, focusing on K, G (our Sun is a G2 star) and F type stars. All stars have a fair distribution of planets, although K stars seem to have more super-Earths, this is probably due it being easier to detect small planets against faint stars.
So what does this mean for the detection of Earth-like planets in Earth-like orbits? It's hard to tell, as the issues of sampling mean that planets with orbital periods like Earth (365 days) will be under represented. And of course some of the detections will be false positives, but that shouldn't distort the statistics too much. Kepler sampled 156,000 stas and found planets around 706 of them, while only 0.5% of the stars observed by Kepler were found to have planets, remember that we can only see transits if the systems are nearly edge on to us, and orbits much longer that 50 days will be likely to be significantly underrepresented in systems that are edge on.
Over at Systemic Greg Laughlin did an interesting calculation, he assumed that half of the stars Kepler looked at had planets with orbital periods of 50 days or less, and that the solar systems orientations were random. He found that Kepler should have seen around 1100 stars with planets, not far from what was actually reported. The question of whether Solar System-like systems are rare or common depends on whether the majority of F, G and K stars have planets, so the incidence of Solar System like systems remains uncertain.
The 400 systems held over for further study will be interesting, the question of whether these are are Earth-like worlds will be answered early in 2011.
The paper discussing the Kepler planets is here, the paper discussing the multiple planet systems is here. You can search the Kepler archive and plot transits or download data here. Here is the Science Daily report, the (somewhat disappointing) Kepler Press release, the Dynamics of Cats report is here, the Systemic report is here and Starts with a Bang's musings are here.
Links to this post: