its own gravity like miniversions of the Sun (Boss, 2002). In this model the excess abundances of heavy elements in Jupiter and Saturn would have been acquired later by capture of smaller rocky and icy bodies. This model, however, does not account well for the compositions of Uranus and Neptune, which do not have very much gas. Other important questions about the outer planets are when they formed and the extent to which they may have drifted inward or outward from the Sun during and after formation. Where the outer planets were and when is important for understanding how the inner planets formed.

The primary difference between the inner and outer planets (rock versus gas and ice) is thought to reflect the temperature gradient in the solar nebula. Temperatures were relatively high (>1000 K) near the developing Sun, dropping steadily with distance. Near the Sun, mainly silicates and metal would have condensed from the gas (so-called refractory materials), whereas beyond the asteroid belt, temperatures were low enough for ices (i.e., water, methane, ammonia) containing more volatile elements to have condensed, as well as solid silicates. It was once thought that as the nebula cooled, solids formed in a simple unidirectional process of condensation. We now know that solids typically were remelted, reevaporated, and recondensed repeatedly as materials were circulated through different temperature regimes and variously affected by nebular shock waves and collisions between solid objects. Important details of the temperatures of the solar nebula, however, are still uncertain, including such significant issues as peak temperatures, how long they were maintained, and how temperature varied with distance from the Sun and from the midplane of the disk. Defining these conditions is an important part of understanding how the chemical compositions of the planets and meteorites came to be.

The standard model for the formation of the inner planets is somewhat more complicated than the model for outer-planet formation and is based largely on theory and anchored in information from meteorites and observations of disks around other stars (Chambers, 2003). The model strives to explain how a dispersed molecular cloud with a small amount of dust could evolve into solid planets with virtually no intervening gas and how the original mix of chemical elements in the molecular cloud was modified during that evolution. Significant unknowns are how long the process took, how solid materials were able to coagulate into progressively larger bodies, and how and when the residual gas was dissipated. The time for centimeter-sized solid objects to form at Earth’s distance from the Sun, according to the standard model, might have been as short as 10,000 years. These small solid objects were highly mobile, pulled Sun-ward large distances by the Sun’s gravity as a result of drag from the still-present H-He gas. Submeter-sized objects were also strongly affected by turbulence in the gas.

A particular deficiency of the standard model is its inability to describe the formation of kilometer-sized bodies from smaller fragments. The current best guess is that the dust grains aggregated slowly at first, and growth accelerated along with object size as small objects were embedded into larger ones (Weidenschilling, 1997). The aggregation behavior of objects greater than a kilometer in size is better understood: they are less affected by the presence of gas than are smaller pieces, and their subsequent evolution is governed by mutual gravitational attractions. Growth of still larger bodies, or planetesimals, from these kilometer-sized pieces should have been more rapid, especially at first. Gravitational interactions gave the largest planetesimals nearly circular and coplanar orbits—the most favorable conditions for sweeping up smaller objects. This led to runaway growth and formation of Moon- to Mars-sized planetary embryos. Growth would have slowed when the supply of small planetesimals was depleted and the embryos evolved onto inclined, elliptical orbits. Dynamical simulations based on statistical methods and specialized computer codes are finding that a number of closely spaced planetary embryos are likely to have formed about 100,000 years after planetesimals appeared in large numbers (e.g., Chambers, 2003).

The later stages of planet formation took much longer, involved progressively fewer objects, and hence are less predictable (Figure 1.4). The main phase of terrestrial planet formation probably took a few tens of millions of years (Chambers, 2004). The final stages were marked by the occasional collision and merger of planetary embryos, which continued until the orbits of the resulting planets separated sufficiently to be protected from additional major collisions.

Although there are four terrestrial planets, models suggest that the number could easily have been three



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement