A method for the direct determination of the surface gravities of transiting extrasolar planets Options

[Mongo]

A method for the direct determination of the surface gravities of transiting extrasolar planets

John Southworth, Peter J. Wheatley, Giles Sams

Abstract: We show that the surface gravity of a transiting extrasolar planet can be calculated from only the spectroscopic orbit of its parent star and the analysis of its transit light curve. This does not require additional constraints, such as are often inferred from theoretical stellar models or model atmospheres. The planet's surface gravity can therefore be measured precisely and from only directly observable quantities. We outline the method and apply it to the case of the first known transiting extrasolar planet, HD 209458b. We find a surface gravity of g_p = 9.28 +/- 0.15 m/s, which is an order of magnitude more precise than the best available measurements of its mass, radius and density. This confirms that the planet has a much lower surface gravity that that predicted by published theoretical models of gas giant planets. We apply our method to all fourteen known transiting extrasolar planets and find a significant correlation between surface gravity and orbital period, which is related to the known correlation between mass and period. This correlation may be the underlying effect as surface gravity is a fundamental parameter in the evaporation of planetary atmospheres.

Summary Results (metres per second squared), from highest to lowest:

28.3 +- 4.4 -- OGLE-TR-113
21.5 +- 3.5 -- HD189733b
20.7 +- 2.6 -- TrES-2
20.1 +- 2.7 -- WASP-2
18.0 +- 6.0 -- OGLE-TR-132
17.9 +- 1.9 -- OGLE-TR-56
16.4 +- 2.5 -- HD149026b
16.1 +- 1.0 -- TrES-1
13.3 +- 4.2 -- OGLE-TR-111
13.3 +- 2.5 -- XO-1
10.6 +- 1.7 -- WASP-1
9.5 +- 2.1 -- OGLE-TR-10 (edited from 4.5 +- 2.1, an apparent typo)
9.28 +- 0.15 -- HD209458b
7.1 +- 1.1 -- HAT-P-1

By comparison, here are the surface gravities of the Solar gas giants:

23.12 -- Jupiter
11.15 -- Neptune
8.96 -- Saturn
8.69 -- Uranus

Bill

[nprev]

10.6 +- 1.7 -- WASP-1
9.28 +- 0.15 -- HD209458b

...therefore, these bodies are close to the mass of Earth?!

[helvick]

Saturn's surface gravity is only about 14% higher than Earth's even though it's 95 times more massive so you cannot infer that that object is earth sized just from the surface gravity. I'm pretty certain that it's not; no earth sized extra-solar planets have been found yet.

[Greg Hullender]

WASP-1 has about 87% the mass of Jupiter, and HD209458b has about 0.66% the mass of Jupiter, so they're both way, way larger than Earth.

They're both among the lighter planets in this group, though, and orbiting stars that hotter than most, so that probably bloats their atmospheres.

At the other end of the scale, OGLE-TR-113 is the very heaviest (1.35x the mass of Jupiter) AND orbiting the very dimmest of the stars.

Here's my source of data: http://obswww.unige.ch/~pont/TRANSITS.htm

From this data, you can estimate the surface gravity, since you've got mass and radius, and the numbers in the paper are mostly pretty close (within error bars), except for OGLE-TR-10, which is 2.25 times lower than expected. Wonder if that was supposed to be 9.5 instead of 4.5.

--Greg

[edstrick]

Nominally, there are 4 primary variables:

Planetary Mass
Stellar irridiation of the planet (watts/square meter)
Planetary age (is it young and still bloated from accretional heating?)
Composition: Hydrogen/Helium vs Carbon/Nitrogen/Oxygen (water) content vs silicates/metal.

A secondary variable is how differentiated the planet is... all the silicates and metal in a core or dispersed through much of the volume.
Mass is directly observed from radial velocity measuremts of the transiting planet's star
Irradiation depends on planetary orbit and stellar output (and planetary reflectance)
Planetary age can be estimated from stellar spectra, but is a pretty low precision estimate (I think) for middling-young planets (?few hundred million years?)

Composition is the one that's really hard to measure. This fascinating bit of analytical work will help a lot with that.. and the resulting constraint on accretion/orbit evolution models.

[Mongo]

From this data, you can estimate the surface gravity, since you've got mass and radius, and the numbers in the paper are mostly pretty close (within error bars), except for OGLE-TR-10, which is 2.25 times lower than expected. Wonder if that was supposed to be 9.5 instead of 4.5.

I wonder too. According to this paper, OGLE-TR-10 has parameters of Mass(Jup.) = 0.63 +- 0.14, and Radius(Jup.) = 1.26 +- 0.07

This gives a range of surface gravities of 6.4 to 12.6 metres per second squared, with a most likely value of about 9.2 metres per second squared.

I will edit the table on the first post of this thread, to account for the apparent typo.

Bill

[Greg Hullender]

Mongo (aka Bill): The 4.5 value is in the original paper, so I suspect the error is theirs. If you look at their charts, it's quite an outlier. Pity, since it makes their correlation claims weaker than they could have been.

It's a neat little paper, actually, and quite accessible to the educated layman. Does anyone here actually know any of the authors? Be nice to help them fix this . . .

--Greg

[John Southworth]

Thanks a lot for your comments on my paper.

I just thought I'd reply to say that the low surface gravity of OGLE-10 (4.5 m/s) is real - it's just a strangely bloated planet. Theorists are still scratching their heads over this one.

There was a new transiting planet found last week (HAT-2), but this one has a huge surface gravity because it's radius is similar to Jupiter's but its mass is eight times larger. An interesting object, but this time in quite good agreement with what theorists have been predicting.

John

[Mongo]

Now that's interesting. So evidently, either the published value for the mass of OGLE-TR-10 is too high, or the published radius is too low (or both), most likely due to the mass of the primary being overestimated by about 100% or the radius of the primary being underestimated by about 30%, or some combination of the two. That's how it looks to me, at any rate.

By the way, glad to see you here! I was fascinated by your paper, and would never have guessed that surface gravity could have been directly determined from the observed data. An good example of original thinking. While I myself am only an interested layman, there are quite a few working scientists who contribute to this board, so you should fit right in.

Bill

[JRehling]

Thanks a lot for your comments on my paper.

I just thought I'd reply to say that the low surface gravity of OGLE-10 (4.5 m/s) is real - it's just a strangely bloated planet. Theorists are still scratching their heads over this one.

FWIW, the extent of atmospheric loss would be a function of escape velocity, not surface gravity. The former is a function of 1/radius, not 1/radius^2. So a big superlow density jovian would still be more like Uranus than Mars in terms of "holding power".

[Greg Hullender]

John: Very pleased to hear from you! So is the error in the "Summary Table of parameters for transiting planets" then?

http://obswww.unige.ch/~pont/TRANSITS.htm

I used their mass and radius values, and got excellent correlation with your paper except for OGLE-10. Agree that it's strangely bloated, but is it THAT strangely bloated?

I thought I cross-referenced it against the original paper to verify the numbers, but I don't have any of that info handy at the moment. Wouldn't be the first time a web page was wrong, after all.

Anyway, when I saw what an outlier it was in your graphs, I figured it'd be worth your while to give it another look; like I said, you get a much better correlation if the other number is correct after all.

--Greg