Pollution of White Dwarfs

White dwarfs are the endpoint of stellar evolution for stars like the Sun and, indeed, for most stars around which we have found planets. Thus, it stands to reason, given the population of planets that have been discovered around other stars, that many white dwarf stars at least had planets at one point in their lifetimes. However, stars lose mass during their evolution and also expand their radii substantially during the giant phases. Thus, some of the close-in planets are likely to be swallowed in this stage and others will move outwards to conserve angular momentum in the face of mass loss. Thus, it is a natural question as to how many planets survive stellar evolution.


A curious datum on this subject is the observation of metal pollution in the atmospheres of many white dwarfs. The current leading model for this phenomenon, championed by my UCLA colleagues Ben Zuckerman and Mike Jura , is that this is the result of accretion of material from rocky asteroids or comets. This is somewhat curious because the red giant stage should have eliminated any rocky material that close to the star, pointing to a post-red giant, dynamical mechanism for the source population. My current graduate student, Shane Frewen , is working on exactly such a mechanism. He has investigated the scattering of asteroids by eccentric planets around both main sequence stars and white dwarfs, to quantify the efficiency with which a planet at several AU can guide asteroids onto nearly radial orbits that strike the central star. He has shown that massive, eccentric planets can provide such a population of asteroids, although the rate of pollution decays with time as the unstable orbits become depopulated. The relevant paper can be found here .

Pulsar Planets

The first confirmed extrasolar planets were actually those discovered around a pulsar in 1992. The fact that earth-like planets can be produced in such an exotic location has been a source of fascination for me and I have revisited the standard theory for planet formation in this modified context with both graduate and undergraduate students (see here and here ).


This study had an interesting feedback on more traditional planet formation theories. In the course of the above work, we arrived at a preferred model in which the pulsar planets resulted from the deposition of a narrow annulus of solid material at the outer edge of an expanding gaseous disk. I noted that this model gave an excellent fit to the observations but also seemed to reproduce many of the features of the terrestrial planets of our own solar system.


The figure at left shows the results of simulations of the formation of the terrestrial planets if we assume that the initial conditions are two earth masses of solids in planetesimals confined initially in a narrow annulus between 0.7 and 1 AU. The open circles are the end products of an ensemble of simulations, while the solid points indicate the masses and locations of Mercury, Venus, Earth and Mars (Mercury has an error bar because it may have lost some mantle material in a late giant collision - a favourite explanation for its high density).


This model provides an explanation for the low masses of Mercury and Mars, namely that they formed from diffusion tails of material scattered out of the original annulus. This has been a problem for more traditional power-law initial conditions as there is no obvious physical process that sets this scale. Additional evidence in favour of this model is that it matches the rapid assembly times for the planets that are inferred from radiogeochemical measurements. The relevant paper is shown here .


The origin of such an annulus is still a matter for debate. One possibility is that the annulus represents the location of a pressure maximum in the disk that trapped smaller migrating models. Another model that has received some attention of late is the so-called `Grand Tack' model, in which Jupiter migrates inwards to 1.5 AU before moving out again, which then sculpts the outer edge of the disk.

The question of how to fit our solar system into the overall context of the extrasolar planets is still very uncertain, and the lack of planets with orbital periods less than 30 days in our solar system is an enduring mystery, which suggests that we are still missing a fundamental piece of the puzzle. This may be related to the unusually larger number of single transit systems mentioned before and may indicate a more violent dynamical history for the solar system than is usually assumed.