Planetary science: Preventing stars from eating their young.

Writing in this issue, Benítez-Llambay et al.1 (page 63) report computer simulations that may throw light on why the Solar System and a substantial fraction of exoplanetary systems have gas-giant planets that are separated from their host stars by distances of at least Earth's distance from the Sun. Gas giants are planets such as Jupiter (Fig. 1) and Saturn, with masses tens to hundreds of times that of Earth and consisting mostly of gas. In the widely accepted model of gas-giant formation (for a review, see ref. 2), an embryonic solid core forms in a disk of gas and dust that orbits a newly forming star. The embryo grows by colliding with, and thereby accreting, smaller solid bodies from the circumstellar disk. If it grows sufficiently quickly, it can reach a critical solid mass that is large enough — about 10 Earth masses — for it to undergo rapid accretion of the surrounding gas and to mature into a gas giant before the disk gas disperses (a process that takes 1 million to 10 million years). For more than 30 years3, 4, it has been recognized that this scenario suffers from a potentially fatal flaw. The growing planetary embryo raises a gravitational wake in the gas disk that is predicted to exert a tidal torque on the planet and make it spiral towards the surface of the star much faster than it can grow. The spiralling occurs within about 100,000 years for an embryo of a few Earth masses. However, roughly 10–20% of exoplanetary systems and, of course, our own Solar System, have gas giants that are beyond 1 astronomical unit (AU) from their host stars5 (1 AU is the median Earth–Sun separation), so nature has apparently overcome this difficulty. Interest in the past few years has focused6, 7 on the gravitational effects of the gas in the annulus around the planet's orbit, where the rotational speed of the gas is close to that of the embryo: the corotation region. In their simulations of a solid planetary embryo orbiting in a gas disk, Benítez-Llambay et al. show that the energy released by the material in solid bodies that are accreting onto the growing embryo can substantially heat the gas in the corotation region near the embryo and cause regions just behind and just ahead of the embryo's orbital path to expand and become under-dense. The region behind expands more than the region ahead (for reasons given below), and so the latter pulls the embryo forward more than the former pulls it back. This results in a positive contribution (which the authors call a “heating torque”) to the embryo's angular momentum that causes the embryo to migrate outward from the parent star and that can, under certain circumstances, overcome the negative contribution from the spiral wake. Just what are those circumstances? The authors first show that the asymmetry in the heating and expansion of regions ahead of and behind the embryo is related to the fact that a parcel of gas in the disk that is at the same distance from the star as the embryo moves slightly more slowly in its circular orbit than the embryo, provided that the gas pressure decreases with increasing distances from the star. The gradient in the gas pressure helps to support the gas parcel in the same way that the vertical gradient in Earth's atmospheric pressure prevents the atmosphere from collapsing on our heads. In the stellar context, it means that the region in the gas disk that exactly corotates with the embryo is within the embryo's orbit. This causes the gas flow and heating to differ ahead of and behind the embryo. Thus, in most regions of the disk the heating torque is robustly positive. However, the magnitude of the heating torque depends on the disk's opacity (which determines how far from the embryo the energy of accreted material is deposited), the rate at which material is accreted (which controls the rate at which energy is liberated and hence is linearly related to the heating torque), and the mass of the embryo. For embryo masses near one Earth mass, for which migration towards the star most threatens gas-giant formation, if the accretion rate doubles the embryo mass within about 60,000 years, the heating torque will overcome the tidal torque for the 'standard' values of the disk properties adopted by Benítez-Llambay and co-workers. Therefore, the results suggest a bifurcation in embryo behaviour: disks in which embryo growth is rapid can go on to produce gas giants at or beyond 1 AU, whereas disks in which accretion of solids is slow may lead to 'failed cores', which migrate inward. The authors note that this bifurcation could explain the strong observed correlation in extrasolar systems between the abundance of heavy elements in the host star (elements, other than hydrogen and helium, that would be in the form of solids to be accreted) and the presence of gas giants beyond 1 AU from it. Benítez-Llambay and colleagues' simulations are just a first attempt at modelling the complex coupling between the processes of hydrodynamics and of energy transfer that take place in the circumstellar-disk regions near a growing planetary embryo. Further modelling is needed, especially in view of studies8, 9 of the high accretion rates associated with pebble-sized objects in the presence of gas disks. Nonetheless, the authors' results underscore what has become increasingly apparent in the past decade — that the amazingly diverse nature of exoplanetary orbits and the structure of the planets on them force us to carefully examine our models of planet formation in an attempt to discern which of many complex and interrelated physical phenomena are likely to be most significant in shaping planetary-system architectures. The present study may well contribute an important ingredient to the mix. Download references Subscribe to comments