≡ Menu

On Kepler’s First Planets

Just how good is Kepler at finding planets? We’re getting a pretty good idea. In his talk yesterday at the AAS meeting in Washington, William Borucki (NASA Ames) showed a plot of the lightcurve for previously known planet HAT-P-7. The signature of the planetary transit is unmistakable in these data, a well defined dip in the starlight as HAT-P-7 makes the star just a little dimmer by its passage. Kepler’s sensitivity is apparent.

But the plot is more fascinating still, for in addition to the well defined signature that denotes the dip in starlight caused by the planet moving across the face of the star, Kepler also saw a second dip. That one was caused by the light of the planet being blocked by the star itself. It’s a tiny dip, but one readily demonstrated in Borucki’s chart, and it tells us that Kepler is living up to expectations in terms of finding faint signals. We all hope, of course, for a future finding, the faint signal of a terrestrial world, preferably one in the habitable zone of a star not so different from our own.

Not that the first Kepler planets weren’t fascinating in their own right. What Borucki announced yesterday were more ‘hot Jupiters,’ planets of high mass and extreme temperatures, the five ranging from something similar to Neptune to larger than Jupiter, with estimated temperatures between 1200 and 1650 degrees Celsius. They’re hot, bright, and extremely close to their stars. Kepler 7b is one of the least dense planets yet discovered.

All of these worlds orbit stars larger and hotter than the Sun, and in that regard all are more or less what we expected in early Kepler results, the low-hanging fruit that turns up readily while smaller planets with longer period orbits take longer to discern and be confirmed. Right now, after all, we’re working with no more than six weeks of data collected since May. Note this from Greg Laughlin’s systemic site:

The Kepler planets are primarily orbiting high-metallicity, slightly inflated, slightly evolved stars. These particular parent stars were likely selected for high-priority confirmation observations because their abundant, narrow spectral lines should permit maximally efficient, cost-effective Doppler-velocity follow-up.

We also have an unusual find, a system where the light curve from the apparent planet dips more strongly during the occultation than the transit, suggesting it is hotter than the parent star. What has astronomers puzzled is that early indications are the object is too big to be a white dwarf, and Borucki noted that this wasn’t the only example Kepler had found of such an object. All in all, we’re seeing good things for Kepler, adding Kepler 4b, 5b, 6b, 7b and 8b to our exoplanet list, and finding unexpected things the explanation of which should further refine our planetary formation models.

You can get a look at numerous preprints for Kepler’s newly revealed work here (this is the arXiv page for new submissions, so it will be changing), and I note especially Latham et al., “Kepler-7b: A Transiting Planet with Unusually Low Density,” citing a world whose mass is less than half that of Jupiter, but whose radius is fifty percent larger (preprint). We’re looking at a density of 0.17 g/cc for a planet orbiting a star considerably larger than the Sun, one presumed to be near the end of its life on the Main Sequence.


Comments on this entry are closed.

  • Horatius January 5, 2010, 11:21

    can someone explain me why only 5 planets where announced? Does it mean that they analyzed only part of their 43days dataset e.g. selected several stars with particulary good characteristics and focused their attention only on them? Or maybe they automatically analyzed whole 100K sample and found only 5 objects orbiting their suns in about 5 days periods (there is no more hot jupiters with such short periodicity in the sample)? They expected previously several hundreds of hot-jupiters to be found in the data – therefore I assume that in fact they found much larger of them in the data but selected only five of them for ground based observations in order to have them confirmed. But who knows…

  • Administrator January 5, 2010, 14:44

    Horatius, here’s the answer, via Greg Laughlin’s site:

    According to a S&T editor Bob Naeye, who reported on Bill Borucki’s scientific talk this morning, the first 43 days of photometric observations from the satellite generated 175 transit candidates, of which 50 were followed up in detail to extract the 5 announced planets. The Keck I telescope has been the major workhorse for the high-precision RV follow-up efforts that are required to get accurate masses. According to the Keck I Telescope Schedule, 17 nights were allocated to the Kepler team from July through December of last year. Within this time alotment, roughly 50 RV measurements for the 5 new planets were obtained. The velocity precision for Kepler-4b looks to be of order 2-3 m/s, which is excellent.

  • andy January 5, 2010, 16:13

    Personally my favourite result that’s announced in today’s arXiv batch is the detection of the tidal bulge on the star HAT-P-7 induced by the orbiting hot Jupiter. Seems that Kepler has the capability to characterise the brighter transiting planets very well indeed.

  • ron smith January 5, 2010, 23:25

    I have 2 questions about the future results from Kepler.

    At some point the planets found by Kepler are going to be impossible to verify by ground based doppler measurements because they are too small. Is some alternate method planed for verifying them?

    I don’t understand why they are talking about only 6 weeks of data. Hasn’t Kepler been commisioned for over 6 months now (since May 09)?

  • Ronald January 6, 2010, 4:31

    I am looking forward to the time that it will be possible to do statistically significant analysis of Kepler data, i.e. answer the question what proportion of various type stars (spectral type, metallicity, etc.) possess various types of planets, particularly from hot (sub) giants to our type of system.

    I understood that, not unlogically, the period of the observed light curve (sinusoid) corresponds with the orbital period of the transiting planet, hence, as time proceeds, ideally it should be possible to plot number of planets against orbital period. What worries me a bit is the need for Keck confirmation, at least for mass, since, according to Greg Laughlin’s systemic site, this is the bottleneck (because of observation time constraints and sensitivity limits).

    I wonder, would it still be possible, even without any ground RV confirmation, to come up with (at least) quantitative data with regard to (number of planets) and orbital period/distance?

  • david lewis January 6, 2010, 7:08

    175 transit candidates? That sounds promising.

    At the radius of those planets from their stars the generals odds of detecting a planet by transit would be? Just to take a guess and say 10 percent then that would imply among the 100,000 stars there would be 1,750 planets (if all 175 transits candidates are planets) in very close orbits.

    However in the solar system we have planets ranging out for billions of kilometers. So the odds of looking at a few million kilometers out of a few billion would imply the chances of detecting a planet in our solar system by looking at such a narrow orbit is maybe 1 in a thousand? With many stars being smaller 1 in 100 or 1 in 200 might be a better average across a large number of star systems. So that 1,750 then becomes 175,000 to 350,000 potential planets?

    And then there are the smaller rocky worlds kepler would not detect.

    Of course we know nothing about the distribution of planets within star systems on average yet. But it is fun to hope.

  • djlactin January 6, 2010, 10:29

    “We also have an unusual find, a system where the light curve from the apparent planet dips more strongly during the occultation than the transit,”
    How can they tell which is the occultation and which is the transit? If they’re wrong, the conclusion is also (very) wrong.

  • andy January 6, 2010, 14:08

    djlactin: I’d guess they are using radial velocity data to determine which is the occultation and which is the transit. Once you have the RVs it should be pretty easy to figure out.

  • Wedge January 7, 2010, 9:55

    @andy @djlactin, radial velocity can provide this evidence but photometry alone would be sufficient. With such a bright planet they would not only be able to detect occultations and transits but also the change in total brightness (star plus planet) as the planet changes phase. This would be enough on its own to determine which event was the occultation and which the transit (the light curve should grow brighter as the planet waxes then is occulted and grow dimmer as the planet wanes then transits in front of the star).

  • Ronald January 7, 2010, 12:43

    Paul, you state:
    “All of these worlds orbit stars larger and hotter than the Sun, and in that regard all are more or less what we expected in early Kepler results, the low-hanging fruit that turns up readily while smaller planets with longer period orbits take longer to discern and be confirmed.”

    I would think that is partly right, i.e. the first part of your statement may be slightly faulty: the ‘low-hanging fruit that turns up readily’ are the large mass *planets* in close orbit aroung their parent star, not necessarily around large (and hot) *stars*, as you say. Or am I missing something here?

    Furthermore, I find it striking that all of these 5 stars are large and hot and mostly (at least 4 out of 5) quite metallic.
    Is that just an observational choice, or does this tell us something about the occurrence of hot (sub)giant planets?

  • andy January 7, 2010, 15:32

    Oops, forgot reflection effects in binary systems, thanks for pointing that out Wedge.

  • Administrator January 7, 2010, 19:09

    Ronald wrote:

    I would think that is partly right, i.e. the first part of your statement may be slightly faulty: the ‘low-hanging fruit that turns up readily’ are the large mass *planets* in close orbit aroung their parent star, not necessarily around large (and hot) *stars*, as you say. Or am I missing something here?

    Probably my text could have been phrased better, but I’m saying the same thing you are, Ronald, so no disagreement.

  • ljk January 10, 2010, 1:35

    Discovery of the Transiting Planet Kepler-5b

    Authors: David G. Koch (NASA Ames Research Center), William J. Borucki
    (NASA Ames Research Center), Jason F. Rowe (NASA Ames Research Center), Natalie M. Batalha (San Jose State University), Timothy M. Brown (Las Cumbres Observatory Global Telescope), Douglas A. Caldwell (NASA Ames Research Center), John Caldwell (York University), William D. Cochran (University of Texas, Austin), Edna DeVore (SETI Institute), Edward W. Dunham (Lowell Observatory, Flagstaff), Andrea K. Dupree (Harvard-Smithsonian Center for Astrophysics), Thomas N. Gautier III (Jet Propulsion Laboratory), John C. Geary (Harvard-Smithsonian Center for Astrophysics), Ron L. Gilliland (Space Telescope Science Institute), Steve B. Howell (National Optical Astronomy Observatory), Jon M. Jenkins (SETI Institute), David W. Latham (Harvard-Smithsonian Center for Astrophysics), Jack J. Lissauer (NASA Ames Research Center), Geoff W. Marcy (University of California, Berkeley), David Morrison (NASA Ames Research Center), Jill Tarter (SETI Institute) et al. (4 additional authors not shown)

    (Submitted on 6 Jan 2010)

    Abstract: We present 44 days of high duty cycle, ultra precise photometry of the 13th magnitude star Kepler-5 (KIC 8191672, Teff=6300 K, logg=4.1), which exhibits periodic transits with a depth of 0.7%. Detailed modeling of the transit is consistent with a planetary companion with an orbital period of 3.548460+/-0.000032 days and a radius of 1.431+/-0.050 Rj.

    Follow-up radial velocity measurements with the Keck HIRES spectrograph on 9 separate nights demonstrate that the planet is more than twice as massive as Jupiter with a mass of 2.114+/-0.057 and a mean density of 0.894+/-0.079 g/cm^3.

    Comments: 13 pages, 3 figures, submitted to the Astrophysical Journal Letters

    Subjects: Earth and Planetary Astrophysics (astro-ph.EP)

    Cite as: arXiv:1001.0913v1 [astro-ph.EP]


  • ljk May 31, 2011, 3:30

    Exoplanet Kepler-7b Unexpectedly Reflective

    by Jon Voisey on May 30, 2011

    Early on in the hunt for extra solar planets, the main method for discovering planets was the radial velocity method in which astronomers would search for the tug of planets on their parent stars. With the launch of NASA’s Kepler mission, the transit method is moving into the spotlight, the radial velocity technique provided an early bias in the detection of planets since it worked most easily at finding massive planets in tight orbits. Such planets are referred to as hot Jupiters. Currently, more than 30 of this class of exoplanet have had the properties of their emission explored, allowing astronomers to build a picture of the atmospheres of such planets. However, one of the new hot Jupiters discovered by the Kepler mission doesn’t fit the picture.

    The consensus on these planets is that they are expected to be rather dark. Infrared observations from Spitzer have shown that these planets emit far more heat than they absorb directly in the infrared forcing astronomers to conclude that visible light and other wavelengths are absorbed and reemitted in the infrared, producing the excess heat and giving rise to equilibrium temperatures over 1,000 K. Since the visible light is so readily absorbed, the planets would be rather dull when compared to their namesake, Jupiter.

    The reflectivity of an object is known as its albedo. It is measured as a percentage where 0 would be no reflected light, and 1 would be perfect reflection. Charcoal has an albedo of 0.04 while fresh snow has an albedo of 0.9. The theoretical models of hot Jupiters place the albedo at or below 0.3, which is similar to Earth’s. Jupiter’s albedo is 0.5 due to clouds of ammonia and water ice in the upper atmosphere. So far, astronomers have placed upper limits on their albedo. Eight of them confirm this prediction, but three of them seem to be more reflective.

    In 2002, it was reported that the albedo for υAnd b was as high as 0.42. This year, astronomers have placed constraints on two more systems. For HD189733 b, astronomers found that this planet actually reflected more light than it absorbed. For Kepler-7b, an albedo of 0.38 has been reported.

    Revisiting this for the latter case, a new paper, slated for publication in an upcoming issue of the Astrophysical Journal, a team of astronomers led by Brice-Olivier Demory of the Massachusetts Institute of Technology confirms that Kepler-7b has an albedo that breaks the expected limit of 0.3 set by theoretical models. However, the new research does not find it to be as high as the earlier study. Instead, they revise the albedo from 0.38 to 0.32.

    To explain this additional flux, the team proposes two models. They suggest that Kepler-7b may be similar to Jupiter in that it may contain high altitude clouds of some sort. Due to the proximity to its parent star, it would not be ice crystals and thus, would not reach as high of an albedo as Jupiter, but preventing the incoming light from reaching lower layers where it could be more effectively trapped would help to increase the overall albedo.

    Full article here: