Formation of Giant PlanetsJack J. LissauerNASA Ames Research CenterDavid J. StevensonCalifornia Institute of TechnologyThe observed properties of giant planets, models of their evolution and observations of pro-toplanetary disks provide constraints on the formation of gas giant planets. The four largestplanets in our Solar System contain considerable quantities of hydrogen and helium; these gassescould not have condensed into solid planetesimals within the protoplanetary disk. Jupiter andSaturn are mostly hydrogen and helium, but have larger abundances of heavier elements thandoes the Sun. Neptune and Uranus are primarily composed of heavier elements. The transit-ing extrasolar planet HD 149026 b, which is slightly more massive than is Saturn, appears tohave comparable amounts of light gases and heavy elements. The other observed transiting ex-oplanets are primarily hydrogen and helium, but may contain supersolar abundances of heavyelements. Spacecraft flybys and observations of satellite orbits provide estimates of the grav-itational moments of the giant planets in our Solar System, which in turn provide informationon the internal distribution of matter within Jupiter, Saturn, Uranus and Neptune. Atmosphericthermal structure and heat flow measurements constrain the interior temperatures of these plan-ets. Internal processes may cause giant planets to become more compositionally differentiatedor alternatively more homogeneous; high-pressure laboratory experiments provide data usefulfor modeling these processes.The preponderance of evidence supports the core nucleated gas accretion model. Accordingto this model, giant planets begin their growth by the accumulation of small solid bodies, asdo terrestrial planets. However, unlike terrestrial planets, the giant planet cores grow massiveenough to accumulate substantial amounts of gas before the protoplanetary disk dissipates.The primary question regarding the core nucleated growth model is under what conditionscan planets develop cores sufficiently massive to accrete gas envelopes within the lifetimes ofgaseous protoplanetary disks.1. INTRODUCTIONThe two largest planets in our Solar System, Jupiter andSaturn, are composed predominantly of hydrogen and he-lium; these two lightest elements also comprise more than10% of the masses of Uranus and Neptune. Moreover, mostextrasolar planets thus far detected are believed (or known)to be gas giants. Helium and molecular hydrogen do notcondense under conditions found in star forming regionsand protoplanetary disks, so giant planets must have ac-cumulated them as gasses. Therefore, giant planets mustform prior to the dissipation of protoplanetary disks. Opti-cally thick dust disks typically survive for only a few mil-lion years (see chapters by Briceno et al. and by Wadhwaet al.), and protoplanetary disks have lost essentially all oftheir gases by the age of < 107years (see chapter by Meyeret al.), implying that giant planets formed on this timescaleor less.Jupiter and Saturn are generally referred to as gas giants,even though their constituents aren’t gasses at the high pres-sures that most of the material in Jupiter and Saturn is sub-jected to. Analogously, Uranus and Neptune are frequentlyreferred to as ice giants, even though the astrophysical icessuch as H2O, CH4, H2S and NH3that models suggest makeup the majority of their mass (Hubbard et al., 1995) arein fluid rather than solid form. Note that whereas H andHe must make up the bulk of Jupiter and Saturn becauseno other elements can have such low densities at plausibletemperatures, it is possible that Uranus and Neptune are pri-marily composed of a mixture of ‘rock’ and H/He.Giant planets dominate our planetary system in mass,and our entire Solar System in angular momentum (con-tained in their orbits). Thus, understanding giant planet for-mation is essential for theories of the origins of terrestrialplanets, and important within the understanding of the gen-eral process of star formation.The giant planets within our Solar System also sup-ported in situ formation of satellite systems. The Galileansatellite system is particularly impressive and may containimportant clues to the last stages of giant planet forma-tion (Pollack and Reynolds, 1974; Canup and Ward, 2002;Mosqueira and Estrada, 2003a, b). Ganymede and Cal-listo are roughly half water ice, and Callisto has most ofthis ice mixed with rock. It follows that conditions mustbe appropriate for the condensation of water ice at the lo-cation where Ganymede formed, and conditions at Callisto1must have allowed formation of that body on a time scaleexceeding about 0.1 million years, so that water ice wouldnot melt and lead to a fully differentiated structure. Themore distant irregular satellite systems of the giant planetsmay provide constraints on gas in the outer reaches of theatmospheres of giant planets (Pollack et al., 1979).The extrasolar planet discoveries of the past decade havevastly expanded our database by increasing the number ofplanets known by more than an order of magnitude. Thedistribution of known extrasolar planets is highly biased to-wards those planets that are most easily detectable using theDoppler radial velocity technique, which has been by far themost effective method of discovering exoplanets. These ex-trasolar planetary systems are quite different from our SolarSystem; however, it is not yet known whether our plane-tary system is the norm, quite atypical, or somewhere inbetween.Nonetheless, some unbiased statistical information canbe distilled from available exoplanet data (Marcy et al.,2004, 2005; chapter by Udry et al.): Roughly 1% of sun-like stars (late F, G and early K spectral class main se-quence stars that are chromospherically-quiet, i.e., have in-active photospheres) have planets more massive than Sat-urn within 0.1 AU. Approximately 7% of sunlike stars haveplanets more massive than Jupiter within 3 AU. Planets or-biting interior to ∼ 0.1 AU, a region where tidal circular-ization timescales are less than stellar ages, have small or-bital eccentricities. The median eccentricity observed forplanets on more distant orbits is 0.25, and some of theseplanets travel on very eccentric orbits. Within 5 AU ofsunlike stars, Jupiter-mass planets are more common thanplanets of several Jupiter masses, and substellar compan-ions that are more than ten times as massive as Jupiter arerare. Stars with higher metallicity are much more likelyto host detectable planets
View Full Document
Unlocking...