Colliding planets and disappearing dust.

The collapse of a molecular cloud sets the starting point for the star and planet formation theory. For low mass objects (lower than 3 solar masses) there are (good?) chances that the final product, after few million years, will be similar to our current solar system.

Let's consider from now on a Sun-like star formation scenario. Molecular clouds usually rotate, during their collapse the angular momentum must be conserved. This leads to the formation of an accretion disk around the star: the so called protoplanetary disk. The disk is composed of both dust and gas, as the natal molecular cloud was. The presence of the dust (solid particles with sizes ranging from few nanometers to millimeters) is actually crucial for the planet formation. Processes like coagulation causes  dust growth, until a planetesimal (kilometer sized object) is formed. Subsequent mass accretion on these big bodies, in less than few million years, leads to planet formation.

Depending on their distance from the central star the planetesimals can then still grow until they accrete an icy envelope (far away), a lot of gas (a bit closer), or just keep growing to the size of the Earth (very close to the star). After this stage, the gas in these system should be gone, either accreted from the forming giant planets (see Jupiter), from the star or photo-evaporated by the stellar radiation. After 10 million years we should  then be left with something similar to our solar system, with little amount of free gas around and maybe some dust leftovers.

Let's see now how astronomers observe the stages listed above. To do so we need to define the SED, which stands for: Spectral Energy Distribution. It is a simple plot where one has in the y-axis the flux or luminosity emitted from these objects and in the x-axis the wavelength at which that radiation is emitted. Let's have a look at this figure:

Evolution of an SED. Class 0: only the dust respons to the obscured star is visible. Class I: the star looks for its way out. Class II: star+disk system. Class III: star only? Maybe a planetary system, with dust leftovers (debris disk).

This is the timeline of what happens.
During the collapse the protostar is not yet visible, because of the surrounding dust that absorbs its radiation very efficiently. We can see that the dust is heated, and becomes observable in the far infra-red (100 microns), these are Class 0 objects. As the matter keeps accreting onto the star, the disk starts to form and the star becomes slightly visible (Class I), again the dust is very powerful in absorbing and re-emitting the stellar radiation. Finally, the disk is formed and the situation is the following: we have a visible star (the light colored black-body) and a surrounding disk that gets heated by the optical radiation and re-emits at longer wavelength (Class II objects). This stage lasts until the planets are formed and the disk gets cleared out. We are at this point left with a star, possibly with some planets orbiting and dust leftovers that still might contribute to the excess emission in the infra-red. To go through these steps few million years (~10) are necessary. The SEDs in reality are not that nice, usually astronomers can only choose few wavelengths at which observe the incoming flux, and place points along the solid lines you see in the figure.

Let's also make one thing clear: the smaller the dust, the easier is to observe it. Small grains are very efficient in absorbing and re-emitting the stellar radiation, as they grow they become less and less visible, basically because they shield themselves. A big km-sized object collects less energy in the first place, moreover part of it is wasted internally, instead of being re-emitted toward us. That's why it is so difficult to observe planets, they are just too faint. But this can be the subject for another post.

What follows is taken from a paper appeared in nature last week, and this is a crude simplification of the paper. A 10 million years old object was observed in the early 80s and then in 2006 and 2008. It showed an SED that hinted for the presence of small dust grains at a distance of about 0.4 AU (1 Astronomical Unit = Sun-Earth distance). Nothing weird so far, this object enters quite well the category of Class III objects, where there might be a planetary system orbiting the star and dust leftovers here and there.

Only, 2 years later, the dust disappeared. Look at the observed SED below, especially at the way triangles and  diamonds go down as times goes by. Nowadays there is basically no more emission (factor ~30 less) from this circumstellar dust, but 2 years ago it was there.

SED: flux versus wavelength. Stellar atmosphere on the left, excess due to the disk from 2 microns on. Only upper limits from 60 microns on, there might be no cold dust in the outer disk.

2 years is a very fast timescale for mass clearing in this environment. I hope I read your mind when I raise the following questions:

1) Why there was dust there in the first place?
2) What happened to the dust, then?

1) We know that the disk stopped accreting. When disks are accreting one can observe the hydrogen (and other elements) emission when the accretion column, sucked out from the inner disk, crashes onto the star. This object just doesn't have it. Also the upper limits in the long wavelength of the SED suggests that there is no outer disk (cold, far away dust), which is typically the thing that last the longest. This small dust then was not travelling around the accretion disk toward the star, but it is just standing there, at least from 1983 until 2008. It has to be the leftover of a powerful collision between two planets/planetesimals. The dinamic in the disk at this later stage is chaotic and very difficult to model. Encounters between planets are not rare and their interaction can completely reshape the system. In such collisions solid body can be partially destroyed, replenishing the disk of small material, which can luckily stand out in the SED.

2) After excluding some technical possible answers the authors are left with two scenarios: collisional avalanche or runaway growth.

• The first scenario works as follows: the very small dust particles, without the protection of the gas, which should not be there, are simply swept away by the stellar radiation. If the density is high enough, while they are expelled from the system, several collisions with bigger dust grains can happen, causing the latter to segregate into smaller dust and follow the stream toward the outer disk. This process can work really efficiently and soon enter the "runaway" regime. It would actually lead to clearing timescales of few years! Actually, at 0.4 AU, it should be a lot less..why is it taking so long then?

• The second process is a sort of continuation of what made the disk evolve in the first place. The dust undergoes accretion onto the star, only at a rate higher than usual. But this process in order to be so fast also needs gas! Indeed the gas is able to drag along the small dust particle toward the star speeding up the whole process toward years timescales. But this disk seems to have no gas, the normal dust accretion should proceed slower! Is the gas also coming from the collision that originates the small dust?

Resuming: the first process might be too fast, the second, too slow. Future work on this object will probably clarify this issue, but let's stop for a moment and think about what we are looking at, because these processes might have also shaped our own solar system (where does the Moon come from?). Planets collide and compete again to get bigger and bigger haunting for their collisional ejecta. The fight for survival seems to start way before life as we know it comes out.

Biblio: Melis, Carl; Zuckerman, B.; Rhee, Joseph H.; Song, Inseok; Murphy, Simon J.; Bessell, Michael S., Rapid disappearance of a warm, dusty circumstellar disk, Nature 2012, eprint arXiv:1207.1162.