PDA

View Full Version : On the formation of "hot Jupiters"



baric
2008-Nov-02, 02:09 AM
I've been doing a bit of study on planetary formation recently and am trying to mentally construct a model to better understand the positioning of various gas giants.

Generally, gas giants form outside a star's snow line and build up mass quickly enough to accumulate a think gaseous atmosphere. This has to be done before the nebular gas is removed from the system.

This explains our system's giants quite well, but doesn't work so well for "hot Jupiters" such as the one orbiting 51 Pegasi.

I'm having trouble accepting that Type II migration can account the transition from "outside the snow line" to "4 day orbit around the star" that we see in this particular system.

It would require the core of the gas giant to form (>5 Earth masses), accumulate its atmosphere and still migrate towards the star before the nebular gas was dissipated. If the migration mechanism is that efficient, then Jupiter's own lack of migration in our own system sticks out like a sore thumb!

At that risk of invoking a deus ex machina, I would like to propose an idea to be shot down by the informed denizens of this board :D

What if the gas giant formed around 51 Pegasi formed according to traditional models, but then the system itself traveled through an interstellar cloud... triggering the migration? The gas giant would then travel inward, gobbling up gas and ejecting terrestial planets along the way. By the time it approached the star, there would be no more gas to fuel the migration, resulting in the incredibly tight orbit we see today.

galacsi
2008-Nov-02, 10:01 AM
Or may be some of the gas giants are born more like stars than planets. i.e they are just very small brown dwarfs.

So 51 Pegasi is a double star system. And Jin our own solar system Jupiter and Saturn are small stars too.

baric
2008-Nov-02, 03:52 PM
Or may be some of the gas giants are born more like stars than planets. i.e they are just very small brown dwarfs.

So 51 Pegasi is a double star system. And Jin our own solar system Jupiter and Saturn are small stars too.

The problem with 51 Pegasi is that the Jovian is so incredibly close to the star. There is no plausible scenario in which it could have formed in situ, whether by core accretion or gravitational collapse.

How it formed is less important than its orbit.

galacsi
2008-Nov-02, 04:03 PM
The problem with 51 Pegasi is that the Jovian is so incredibly close to the star. There is no plausible scenario in which it could have formed in situ, whether by core accretion or gravitational collapse.

How it formed is less important than its orbit.

So why had it not formed like a star ? Double stars do exist no ? And with very different configurations.And after all planet , star it is just a name we put on a thing deducted from very small perturbations.

baric
2008-Nov-02, 04:31 PM
So why had it not formed like a star ? Double stars do exist no ? And with very different configurations.And after all planet , star it is just a name we put on a thing deducted from very small perturbations.

The problem is that the planet is just 1/20th of an AU away from the star. Any gas that close to the host star during formation would have either collapsed onto the star itself or been blown away long before the planet could form to a gas giant stage.

galacsi
2008-Nov-02, 08:13 PM
The problem is that the planet is just 1/20th of an AU away from the star. Any gas that close to the host star during formation would have either collapsed onto the star itself or been blown away long before the planet could form to a gas giant stage.

OK I understand your point . But then how do you explain double stars when they are very close of each other ? The same reasoning should apply , don't you think ?

So may be these big planets are not actual planets but stars made in the same process as the main central one.

Just an idea. Don't know how to falsify it.

baric
2008-Nov-02, 11:07 PM
OK I understand your point . But then how do you explain double stars when they are very close of each other ? The same reasoning should apply , don't you think ?

How close is "very close"? A lot of double stars are often many AU apart.

Is there an example of a binary stars with a 4-day orbital period?

timb
2008-Nov-03, 12:13 AM
The problem is that the planet is just 1/20th of an AU away from the star. Any gas that close to the host star during formation would have either collapsed onto the star itself or been blown away long before the planet could form to a gas giant stage.

The standard theory is that the planet migrated to the inner edge of the disk, at which the point the drag stopped and so did the migration.

AndreasJ
2008-Nov-03, 11:03 AM
How close is "very close"? A lot of double stars are often many AU apart.

Is there an example of a binary stars with a 4-day orbital period?

The orbital period of Algol A and B is somewhat less than three days, with a semimajor axis of 0.062 AU. I expect there are even tighter ones.

Binary pulsars are known with orbital periods of a few hours.

timb
2008-Nov-03, 11:31 AM
The orbital period of Algol A and B is somewhat less than three days, with a semimajor axis of 0.062 AU. I expect there are even tighter ones.

Binary pulsars are known with orbital periods of a few hours.

How do they get so close? their stellar progenitors could not have been so close or they would have already collided.

galacsi
2008-Nov-03, 05:10 PM
How close is "very close"? A lot of double stars are often many AU apart.

Is there an example of a binary stars with a 4-day orbital period?

Yes there are every orbital period you want ! You find short orbital period within the class of Spectroscopic binaries.

From wikipedia :
Spectroscopic binaries

Sometimes, the only evidence of a binary star comes from the Doppler effect on its emitted light. In these cases, the binary consists of a pair of stars where the spectral lines in the light from each one shifts first toward the blue, then toward the red, as each moves first toward us, and then away from us, during its motion about their common center of mass, with the period of their common orbit.

In these systems, the separation between the stars is usually very small, and the orbital velocity very high. Unless the plane of the orbit happens to be perpendicular to the line of sight, the orbital velocities will have components in the line of sight and the observed radial velocity of the system will vary periodically. Since radial velocity can be measured with a spectrometer by observing the Doppler shift of the stars' spectral lines, the binaries detected in this manner are known as spectroscopic binaries. Most of these cannot be resolved as a visual binary, even with telescopes of the highest existing resolving power.



Cheers

galacsi
2008-Nov-03, 05:14 PM
How do they get so close? their stellar progenitors could not have been so close or they would have already collided.

They probably collide and exchange some matter before attaining there actual size.

I don't think this is well understood. But the fact is you have very near binaries.

FTL_Diesel
2008-Nov-07, 06:16 PM
So why had it not formed like a star ? Double stars do exist no ? And with very different configurations.And after all planet , star it is just a name we put on a thing deducted from very small perturbations.

We know that Hot Jupiters form differently from stars because their densities (as measured through planetary transits) are much too high for them to have formed through gravitational collapse of the proto-stellar nebula. All Hot Jupiters appear to have compositions that have a *much* higher amount of rock / ice / metals then stars, which indicates that they formed by some form of core-accretion process.

As to the question posed by the OP, migration using the proto-planetary disk is, for right now, the only plausible scenario for moving Hot Jupiters inwards from the snow-line. The ISM (even inside nebulae) is too diffuse to have a substantial impact on planetary orbits.

That being said, why Jupiter didn't migrate is an excellent question, which no one seems to have a good answer for. Perhaps Jupiter just formed late? Maybe there were multiple rounds of planet formation? Whatever the reason, it does seem like migration either "takes" or it doesn't; this is evidenced by the dip in the period distribution from 10 days out to about a year.

baric
2008-Nov-07, 07:27 PM
As to the question posed by the OP, migration using the proto-planetary disk is, for right now, the only plausible scenario for moving Hot Jupiters inwards from the snow-line. The ISM (even inside nebulae) is too diffuse to have a substantial impact on planetary orbits.

This is what I was unaware of. I've been looking for information on this, but with no luck.


That being said, why Jupiter didn't migrate is an excellent question, which no one seems to have a good answer for. Perhaps Jupiter just formed late? Maybe there were multiple rounds of planet formation? Whatever the reason, it does seem like migration either "takes" or it doesn't; this is evidenced by the dip in the period distribution from 10 days out to about a year.

You know, I was pondering this more afterI made the original post. Is it possible that the mass buildup at the snow line serves as an effective barrier for Type II migration (explaining our Jupiter), and that consequently "hot Jupiters" begin *inside* the snow line?

It would seem to require a much more rapid accretion process, so maybe these planets are more of a symptom of higher-density nebulaes.

galacsi
2008-Nov-08, 08:55 AM
We know that Hot Jupiters form differently from stars because their densities (as measured through planetary transits) are much too high for them to have formed through gravitational collapse of the proto-stellar nebula. All Hot Jupiters appear to have compositions that have a *much* higher amount of rock / ice / metals then stars, which indicates that they formed by some form of core-accretion process.

As to the question posed by the OP, migration using the proto-planetary disk is, for right now, the only plausible scenario for moving Hot Jupiters inwards from the snow-line. The ISM (even inside nebulae) is too diffuse to have a substantial impact on planetary orbits.

That being said, why Jupiter didn't migrate is an excellent question, which no one seems to have a good answer for. Perhaps Jupiter just formed late? Maybe there were multiple rounds of planet formation? Whatever the reason, it does seem like migration either "takes" or it doesn't; this is evidenced by the dip in the period distribution from 10 days out to about a year.

What you say make sense.

But ,frankly speaking I am not completely convinced.

Big planets heavier than Jupiter are supposed to be more dense and not bigger. And who knows exactly how stars form ? (Gravitational collapse uh ? , but I am a little prejudiced here)

Thanks for your answer.

timb
2008-Nov-08, 11:35 AM
We know that Hot Jupiters form differently from stars because their densities (as measured through planetary transits) are much too high for them to have formed through gravitational collapse of the proto-stellar nebula. All Hot Jupiters appear to have compositions that have a *much* higher amount of rock / ice / metals then stars,


There are many hot Jupiters of whose composition we know little, so I don't see how you can make that claim. Is Low, C. & Lynden-Bell, D. 1976, MNRAS, relevant? They found the minimum fragment mass in a dark molecular cloud is about 7 MJ. I presume stellar companions smaller than this form by core accretion.


As to the question posed by the OP, migration using the proto-planetary disk is, for right now, the only plausible scenario for moving Hot Jupiters inwards from the snow-line. The ISM (even inside nebulae) is too diffuse to have a substantial impact on planetary orbits.

That being said, why Jupiter didn't migrate is an excellent question, which no one seems to have a good answer for. Perhaps Jupiter just formed late? Maybe there were multiple rounds of planet formation? Whatever the reason, it does seem like migration either "takes" or it doesn't; this is evidenced by the dip in the period distribution from 10 days out to about a year.

It's a dip not a void, and it is partly explainable by observation bias. Radial velocity measurements (and transits) detect close planets more easily than distant ones, astrometry detects distant planets more easily than close ones. Planets in between are hard to detect.

It could be that our type of stellar system is rare. In Gas Disks to Gas Giants: Simulating the Birth of Planetary Systems (arXiv:0808.1439v1) the authors performed a large number of simulations over a range of plausible values for the starting properties of the protoplanetary disk, and found that systems like the solar system arose only for a narrow range of parameters.

FTL_Diesel
2008-Nov-12, 10:14 PM
There are many hot Jupiters of whose composition we know little, so I don't see how you can make that claim.

True, only 52 out of 322 extrasolar planets are transiting, and we only know the densities for these planets. Nevertheless, *all* of the transiting planets have a higher density then the proto-stellar nebula, which is why I feel that claim is pretty sound.


Is Low, C. & Lynden-Bell, D. 1976, MNRAS, relevant? They found the minimum fragment mass in a dark molecular cloud is about 7 MJ. I presume stellar companions smaller than this form by core accretion.

I'm not familiar with that paper, but I think partly yes, and partly no. In free space I'd be inclined to agree with that, but in a proto-stellar environment there will be some substantial gravitational gradient from the cloud that's collapsing into the primary. I'd wager that would increase the minimum mass an object could have if it formed through gravitational collapse, since you'd have to keep everything inside the secondary cloud's Roche Lobe.


It's a dip not a void, and it is partly explainable by observation bias. Radial velocity measurements (and transits) detect close planets more easily than distant ones, astrometry detects distant planets more easily than close ones. Planets in between are hard to detect.

At this point we have a long enough time baseline from the radial velocity surveys that observational bias at periods of less than a year is getting pretty small / dealable. I'd check out Cumming, et. al. (2008) ("The Keck Planet Search: Detectability and the Minimum Mass and Orbital Period Distribution of Extrasolar Planets") for a great paper on the subject.


It could be that our type of stellar system is rare. In Gas Disks to Gas Giants: Simulating the Birth of Planetary Systems (arXiv:0808.1439v1) the authors performed a large number of simulations over a range of plausible values for the starting properties of the protoplanetary disk, and found that systems like the solar system arose only for a narrow range of parameters.

Yeah, it'll be interesting to see if the RV surveys start announcing lots of Solar System-analogues in the next few years, now that they've nearly got enough data to close an orbit like Jupiter's.

FTL_Diesel
2008-Nov-12, 10:23 PM
Is it possible that the mass buildup at the snow line serves as an effective barrier for Type II migration (explaining our Jupiter), and that consequently "hot Jupiters" begin *inside* the snow line?

It would seem to require a much more rapid accretion process, so maybe these planets are more of a symptom of higher-density nebulaes.

Actually, higher density nebulae are supposed to couple more strongly to giant planets, and so migration should be *more* effective (which is supported by some tenative observational evidence). And I'm about 99% sure that anywhere inside the snow line there is not enough mass (within a reasonable accretion radius) to form a planet the size of Jupiter.


Big planets heavier than Jupiter are supposed to be more dense and not bigger. And who knows exactly how stars form ?

That's true, and some of the more massive "exoplanets" (like CoRoT-Exo-3b) are probably the tail end of the star formation process. But for objects within a couple of masses of Jupiter, they're all observed to be too dense to have been formed like stars.

And we actually think we have a pretty good idea of how stars form! At the risk of sounding disreputable, I'd say go read the wikipedia article on star formation as a good jumping off point.

timb
2008-Nov-15, 01:59 AM
Actually, higher density nebulae are supposed to couple more strongly to giant planets, and so migration should be *more* effective (which is supported by some tenative observational evidence). And I'm about 99% sure that anywhere inside the snow line there is not enough mass (within a reasonable accretion radius) to form a planet the size of Jupiter.



That's true, and some of the more massive "exoplanets" (like CoRoT-Exo-3b) are probably the tail end of the star formation process. But for objects within a couple of masses of Jupiter, they're all observed to be too dense to have been formed like stars.


Do you have a reference? I recall reading a paper which I seem to have omitted to file that analysed the masses and radii of hot Jupiters. It had been thought that some hot Jupiters were unexpectedly low density, but the author concluded that their parameters were consistent with a Jovian type composition once you took into account the fact that they were young and, uh, very hot. According to his analysis some of the hot jupiters were 30%+ "metals" as I recall.