Why Do Hot Jupiters Exist and Not Other Planetary Configurations?

[marcus]

Exoplanet statistics have been accumulating since 1995 (and especially since january 1996)

the statistics are observationally biased. it is easier to detect hot jupiters.

but even allowing for the bias, the picture is very strange

massive planets in near circular orbits at jupiter-distance are rare

what is more common is eccentric (oval) orbits and massive planets getting cooked closer even than mercury-distance.

other systems don't seem to sort themselves out into nice circular stable configurations analogous to us. It represents a physical puzzle.
what are the current guesses about why?

 

[marcus]

Please forget habitability issues in this thread. they are jumping the gun and are distracting. Leave your "prevalence of life in the universe" speculations at the door.

this is a more basic physical problem of how do the informed people go about explaining the weird-looking statistics. it is about orbital mechanics and the formation and makeup of these observed planetary systems.

I haven't followed this puzzle but it has been subliminally bugging me for almost 10 years now. Can anybody offer a simple explanation for all the hot jupiters

 

[Gold Barz]

Maybe if we come up with a technique other than the wobble method which mainly focuses on Hot Jupiters wobbling their stars up we will find solar systems with their Jupiters much farther away?

 

[SpaceTiger]

Maybe if we come up with a technique other than the wobble method which mainly focuses on Hot Jupiters wobbling their stars up we will find solar systems with their Jupiters much farther away?

Good, you're thinking like a scientist.

In fact, we are exploring (or rather, have explored) other techniques for planet detection. The first planet actually wasn't discovered by the usual radial velocity technique, but rather by timing the arrival of pulses from a pulsar (a rapidly rotating neutron star). The method is similar to the radial velocity method in that it uses the motion of the star due to the planets, but in this case we're measuring light arrival times rather than spectral shifts. This is an extremely effective technique for planet detection because we can measure the time arrival of pulses very precisely, so much so that we were able to detect a planet of 0.015 M_earth around PSR 1257+12. That's even smaller than Mercury!

Unfortunately, pulsars are rare and these systems bear no resemblance to our own (trust me, you would not want the sun replaced by a pulsar). Another method of planet detection around normal stars is to look for eclipses (or "transits") by the planet. That is, when the planet passes in front of the star, it blocks some of the light, making the star a little bit dimmer. We can detect this dimming with our telescopes and measure how it changes with time to estimate the orbital parameters of the planet. The primary problem with this method is that it requires the planetary system to be inclined directly perpendicular to our line of sight, something that happens in only a small fraction of systems. Until recently, the only really convincing detection was a roughly Jupiter mass object around HD 209458, but the OGLE experiment has been stepping up the detection rate of late.

Also worth noting briefly is that the wobble of the host star exists in both the radial direction (as determined by spectra) and tangential directions (as determined by finding its position on the sky). The latter type of measurement requires extremely high precision "astrometry", something which we hope to achieve with SIM. This should extend the range of massive planets (gas giants, mostly) that we can detect.

What we would really like to do eventually is image the planets directly. This would allow us to compile more information, most importantly its spectrum. The primary difficulty with doing this is that the planets are usually completely obscured by the light of their host star. In the optical, typical planet-star brightness ratios are on the order of one part in a billion. There are missions being planned (most notably TPF) in which they hope to make these images by one of the two following methods:

1) Extremely high resolution imaging in the infrared. This is somewhat easier than the optical because the planet-star brightness ratio is much lower, but it's still a challenging problem that involves deploying a space interferometer.
2) Fancy optics that distort the image of the star to reveal the tiny contribution of light from the planet. This is a bit hard to explain in layman's terms, but the basic idea is that you funnel the light arriving from the central star into a different part of the detector than the light arriving from its outskirts. This may allow us to pick out the dim planet without having to resolve it separately from the star (that is, their images would be partially superimposed).

Unfortunately, such a mission is still a long ways off, so don't expect any ground-breaking detections of earth-like planets in the near future. However, the range of larger planets we can detect is ever-expanding and we should soon be able to give a better idea of how evenly distributed planets are from their host star. In addition to higher-precision instruments to detect less massive planets, we're also compiling a longer time baseline, allowing detection of planets with longer orbital periods. A planet at the size and distance of Jupiter has only just recently become detectable by even the radial velocity method.

 

[marcus]

Let's look at the curve on this page

http://exoplanets.org/

this curve gives some idea of statistics of known planetary systems

do either of you Goldbars or Tiger (pleas I hope) have other links to graphic display of statistics like this?

 

[turbo]

2) Fancy optics that distort the image of the star to reveal the tiny contribution of light from the planet. This is a bit hard to explain in layman's terms, but the basic idea is that you funnel the light arriving from the central star into a different part of the detector than the light arriving from its outskirts. This may allow us to pick out the dim planet without having to resolve it separately from the star (that is, their images would be partially superimposed).

There are optical tricks to improve contrast, one of which can be used by people here with amateur-sized telescopes. I used it specifically to view the faint companion of Sirius after reading of the technique in a popular astronomy magazine (can't remember which one).

The trick is to get out a compass and rule and (using a piece of heavy black posterboard) construct a hexagon just smaller than the aperture of your telescope and surround it with a circle the diameter of the inside of your 'scope's dew shield. Cut out the hexagon, then trim the paper mask so it will fit inside your scope's dew shield. Normally when you view Sirius in an amateur-sized scope, the bright Airey disk of Sirius overwhelms the light of the companion. This mask, however distorts the Airey disk, and gives the image of Sirius (and other star images of tight doubles that you might want to split) six bright spikes with darker zones between them. If you carefully rotate the mask so that Sirius B falls in one of the attenuated zones, you can see it in an amateur scope.

Here is the NASA link for the Orion project, explaining how masking (apodization) might make direct detection of exosolar planets possible.

http://history.nasa.gov/SP-436/ch3.htm

 

[marcus]

I see here 136 planets as of Febrary 2005

http://exoplanets.org/almanacframe.html

the periods of these planets are listed in days

67 of these planets are period less than 332 days
and their axis is less than 1 AU

it is big planets and more than half are "inwards of the earth"

 

[marcus]

here is a good graphic

http://exoplanets.org/massradiiframe.html

 

[Andromeda321]

I was alway under the assumption that because it only takes a few days to detect the wobble of a hot Jupiter they were easier to discover. On the other hand, a planet as far out as Jupiter would take twelve years to orbit once, which means twelve years to see the full wobble. As a result all the planets that take that long an amount of time to orbit once are just being discovered.

 

[SpaceTiger]

If I may be allowed a brief and partially anecdotal response to these issues. I think the assertion that it's just a selection effect is basically correct, but it depends on how you look at the problem. Although it's not clear that there's a peak in the distribution in the "hot Jupiter" regime, there are certainly more than we expected before looking. However, this was mostly due to a pre-Copernicanesque bias that other planetary systems ought to be like ours. From what I've heard, the distribution of planetary radii is roughly uniform in log space (so far), a trend very similar to that seen in binary star systems. However, my roommate is more of the planet guy, so I'll ask him and get back to you.

On the theoretical side, we initially had a hard time explaining "hot Jupiters" because we expected massive planets to form out of light gases at large radii from the star. Although this may still be true, we discovered in our models of protoplanetary disks that the planets ought to "migrate" inwards towards the star as time goes on. This model was so effective at explaining the hot Jupiters that we're now having a hard time explaining how planets can exist that aren't so close to the star.

All in all, I'd say the situation is still a bit of a mess. Let me get back to you in five years.

 

[Gold Barz]

Do you guys believe that there is a fair share of solar system that resembles our own?

 

[Chronos]

This is probably the most complete source of data for exosolar planets:
http://www.obspm.fr/encycl/encycl.html
Vital statistics are given in the Catalog
http://cfa-www.harvard.edu/planets/catalog.html

 

[marcus]

On the theoretical side, we initially had a hard time explaining "hot Jupiters" because we expected massive planets to form out of light gases at large radii from the star. Although this may still be true, we discovered in our models of protoplanetary disks that the planets ought to "migrate" inwards towards the star as time goes on. This model was so effective at explaining the hot Jupiters that we're now having a hard time explaining how planets can exist that aren't so close to the star.

that is interesting.

what dissipates the energy allowing the giants to migrate in?

isnt the whole protoplanetary disk of crud whirling around so why would there be friction, and why wasnt it anticipated?

its really intriguing
=============

... this was mostly due to a pre-Copernicanesque bias that other planetary systems ought to be like ours...

From what I've heard, the distribution of planetary radii is roughly uniform in log space (so far), a trend very similar to that seen in binary star systems.

why SHOULD the log of orbit radii be uniformly distributed?

with binary stars you indicate that the log of the separation of the stars is roungly uniformly distributed over some reasonable range (too close and they touch, too far and other stars would interfere and it wouldn't be stable, so there has to be some bounds)

why should the log of binary's orbit radius be uniform distributed?

you seem to have an interesting life (both now and ahead of you)

looking forward to hearing what yr roommate the planet guy has to say