Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
3 Present Understanding of the Origin of Planetary Systems INTRODUCTION This section gives an overview of current theory regarding fo~ation of planetary systems, with emphasis on relationships with astronomical ob- se~vations. A great deal of effort has been devoted to the complicated problem of describing the formation of a planetary system. What is de- sired is the integration of a wide variety of observational evidence into a theoretical picture that includes a broad range of physical and chemical processes. Development of the theory requires calculations based on the three-dimensional hydrodynamics of a self-gravitating fluid, including the effects of pressure, viscosity, rotation, magnetic fields, shock waves, and tides. Coupled with the hydrodynamic problem are the thermodynamics of the gas and the energy transport through it, by either radiation or convec- tion. Further, one must consider chemical processes, such as the formation of molecules and the formation, growth, and destruction of dust grains, along with their interaction with the gas. Collisions of the dust particles and their accretion into subplanetary objects, as well as gravitational and electromagnetic interactions in a many-body system, also must be incorpo- rated. 'Ib understand the evolution of the central star in a planetary system requires the addition of nuclear physics to the above processes. In the past 200 yr, numerous theories have been put forward regarding the origin of our solar system. Many of the earlier ideas, such as those involving capture of material from the interstellar gas by the Sun or ejection 21
22 of matter from the Sun as a result of a close encounter with a passing star, have been rejected on physical grounds. The classical hypothesis of Emmanuel Kant and Pierre Simon de Laplace, that the planets originated in a disklike nebula surrounding the protosun, forms the basis for most current theoretical work on the problem. The nebula is a by-product of the stellar formation process; that is, the planets and the star are all about the same age. The committee concentrates here on the description of the problem based on the nebular hypothesis, which accounts for many, but by no means all, of the observed facts. Of course, it is possible that other planetary systems are quite different from ours in their orbital and physical characteristics; clearly the favored theory has been strongly influenced by the properties of our own system. As more information becomes available about other systems, substantial modifications of our theoretical ideas will undoubtedly be required, leading to new generalizations not yet envisioned. Four types of observational data are crucial to understanding and potentially solving the general problem of planetary system formation: (1) general dynamical properties of our own planetary system, and statistics and properties of extrasolar planets; (2) properties of regions of current star formation; (3) statistics of multiple stellar systems; and (4) laboratory and spacecraft studies of available solar system materials (meteoritic, cometary, lunar, and terrestrial). The observed dynamical regularity of our system the near coplanarity of the orbits, the small eccentricities and inclinations, the regular spacing of the planets, and the existence of satellite systems with similar regulanties- is probably its most striking property. The masses and compositions of the inner planets as compared with those of the outer planets, as well as the existence of the asteroid belt and the composition and orbital configurations of comets, provide important clues to the nature of the solar nebula and the planetary formation process. The ordered variation in the properties of planets and satellites with radial distance from the Sun is consistent with the interpretation that they are spatially separated samples of an ongmal continuous nebula, although accretion of each body probably occurred over a range of radial distances and the temporal sequence of formation of the planets may not coincide with their present spatial ordering. In this regard it is of great importance to measure orbital inclinations, masses, eccentricities, and other structural properties of extrasolar planetary systems. The second type of observational information, that which describes molecular clouds and stars in the process of formation and in their early history, is extensive and diverse. It includes, for example, radio and infrared measurements. The problem is to identity those particular techniques and data that could provide clues to the origin of planetary systems. Obser- vations of molecular clouds give some indication of the initial conditions for star formation; studies of young objects suggest the presence of disks,
23 dense dust clouds, and mass outflows. The high spatial resolution needed to examine a nebula the size of our present planetary system even in the nearest star-forming regions (for example, at 100 parsecs, 10 AU subtends 0.1 arcsec) has been one of the greatest obstacles to progress In this area. The third type of information relates to stellar multiplicity. A large fraction of all nearby stars of the solar type, perhaps as many as 70 to 90 percent, are members of multiple systems. There are, however, some single stars like the Sun. Whether this fraction changes significantly as a function of the mass of the primary star is observat~onally not well established. An important goal of the theory of star formation, therefore, is to understand the relationship between planetary formation and multiple star formation. In particular, can both occur in the same system? Finally, data from primitive meteorites (and, ultimately, from comets) provide accurate abundance and isotope ratios for many of the chemical el- ements, as well as information regarding pressures and temperatures during the formation phase of these objects and therefore, presumably, of some types of preplaneta~y material. Measurements of both solid material and volatiles inert gases and others in the carbonaceous chondntes provide particularly relevant data. Some information on the magnetic fields in the primordial solar nebula can also be obtained from the meteorites, as can the time constants for some of the principal physical processes. Four major astrophysical processes must be considered, in a unified manner, in the attempt to explain the origin of planetary systems: (1) collapse and star formation in gas and dust clouds; (2) formation, evolution, and dispersal of the disldike nebula; (3) the evolution of the central star; and (43 accretion of the nebular matter into protoplanets. The following sections discuss briefly the state of knowledge on these problems. STAR FORMATION The observational evidence indicates strongly that most if not all star formation takes place in molecular clouds (mean density 10-2~ g cm~3), and probably in the cores of such clouds, where densities are approximately 10-~9 g cm~3 and temperatures about 10 K The basic condition that has to be satisfied for gravitational collapse to occur is the Jeans criterion the requirement that the thermal energy of a volume of gas be less than the absolute value of the gravitational energy. 1b satisfy the Jeans criterion in molecular cloud cores at this density and temperature, about 4 Me of interstellar gas are required. Compression to the required densities could, for example, be initiated by the passage of a shock wave through a cloud; the origin of such shocks could be supernova explosions, an expanding ionized region around an existing hot star, or the shocks associated with spiral density waves in the galaxy. However, the relatively high density
26 density and angular momentum in the cloud just before protostar collapse begins, as well as on the total amount of angular momentum. Note that during the formation of the disk the whole inner structure is well shielded by the du~st-r~ch outer layers of the cloud that are still in the process of collapse. Therefore at this stage the disk may not be directly observable except at radio wavelengths, but its presence could affect the infrared radiation emitted by the protosteller system. The observable layer gradually increases in surface temperature and luminosity as the central mass grows by accretion from the disk and the optical thickness of the collapsing envelope decreases. The central object becomes observable as a so-called T Tauri star, and the protostellar evolutionary phase is completed. The energetic stellar winds and bipolar outflows Netlike collimated flows) observed during the ~105 to 106 yr of the ensuing contraction phase are now widely believed to occur simultaneously with infall of material from the protostellar cloud to the circumstellar disk and inflow of material to the central mass through the dish There is growing evidence that these winds require the presence of an accretion disk and that the mass inflow rate through the disk and the mass outflow rate in the wind are related. It is important to realize that the sequence of events outlined above is simply a sketch and that only a few aspects of protostellar evolution have been calculated in detail, usually with restrictive assumptions. The general problem of collapse of a rotating, magnetic cloud, induding the effects of heating, cooling, ionization, chemistry, and radiation transport, has not been solved. PROPERTIES OF THE NEBULAR DISK If a nebular disk forms as discussed above, it still does not have the proper angular momentum distribution to be in agreement with that deduced observationally for systems consisting of a young star plus a sur- rounding disk. Angular momentum must be transported out from the central regions. Various mechanisms for transport of angular momentum are under study, including gravitational or magnetic torques, viscous effects arising from turbulence or sound waves, or magnetic braking of the cen- tral star both before it appears as an optically visible pre-ma~n-sequence star and after it reaches the main sequence. (Recent work suggests that stellar winds during intervening pre-main sequence stages are ineffective in removing stellar angular momentum.) The importance of these effects is usually studied by means of an idealized model in which the disk Is thin, is small in mass in comparison with the star, is in hydrostatic equilibrium and in Keplerian rotation, and has a temperature that is low enough (<2000 K) that hydrogen is in the molecular form and dust grains are present.
27 Angular momentum transport by turbulence has been studied exten- s~vely, and at least two mechanisms have been suggested for inducing it. First, while matter from the cloud is still collapsing onto the disk, the difference in angular momentum between the existing disk material and the newly accreted material leads to shearing motions that could induce turbulence. Second, after the gravitational infall has stopped, convective instability can be induced by the increase in the opacity of the grains as a function of increasing temperature. The evolution of the system is therefore likely to be that of a viscous disk often known as an accretion disk. The turbulence has two effects: it results in an efficient transfer of heat from the midplane to the surface of the disk, and the Keplerian shear combined with the viscosity due to turbulence induces transfer of mass as well as angular momentum in the dish The dissipation of energy by viscosity provides heat, which is radiated from the surface of the disk. The ultimate energy source for the turbulence and the heat is the gravitational energy of the disk and star interaction. Most of the disk mass eventually sinks slowly toward the central star. At the same time, the angular momentum, along with a small fraction of the mass, is slowly transferred outward. Given an infinite time for evolution, almost all of the nebula would spiral into the central star. However, time scales for circumstellar disk evolution of only a few million years are inferred from theoretical calculations, as well as from recent observations of excess infrared radiation arising from dust embedded in the disks with large infrared excesses at stellar ages of <3 million yr, but fewer than 10 percent of such stars still display this signature by the tune they reach ages of 10 million yr. Such short time scales provide a strong constraint on theories of the planetary formation process, particularly for the gas-rich outer planets. There has been considerable progress during the past few years in our understanding of the evolution of such disks. Yet major uncertainties remain in the theory of convection and turbulence that could affect the deduced evolutionary time scale. It is probable that the other mechanisms for angular momentum transport listed above could have a significant effect during the various phases of disk evolution. The material in the nebular disk that does not condense into planetes- imals or protoplanets is cleared away, arguably only a few million years or less after the formation of the star. A number of mechanisms have been proposed for accomplishing this clearing. (1) Strong stellar winds probably have sufficient energy to sweep out a moderate-mass nebula. But since massive T Taun winds appear to be present only if the star is surrounded by a thick circumstellar disk, and are not seen In stars with low-mass, optically thin disks, it is not obvious what role, if any, is played by such winds in directly sweeping a disk away. (2) Particles in the nebula could be photoionized by ultraviolet radiation from the central star. The extra
29 infrared excesses characteristic of the classical T Taun stars to the small or undetectable infrared excesses of the so-called naked T Tauri stars. THE INVOLUTION OF THE CENTRAL STAR No discussion of the origin of a planetary system is complete without consideration of the central star. The star influences the planetary forma- tion process in several ways. First, its mass as a function of time influences the propertiesfor example, the vertical thicknessof the nebular disk. Second, the mass outflow from the star as it settles into the stellar state may terminate the collapse of the protostellar cloud. The observed bipolar outflows near young stars may be a manifestation of such a process. Some of the infrared objects that exhibit bipolar flow also may be interpreted as having associated disks or tori, with the plane of the disk perpendicular to the flow. As noted above, outflow from the boundary layer and ionization by ultraviolet photons may control the dissipation of the dish Third, the tidal influence of the star plays an important role in the planetary formation process by limiting the region of gravitational influence of the protoplanets. In turn, the evolution of the nebula influences the evolution of the star, as viscous processes transfer matter to the star. The transfer of angular mo- mentum between the star and the nebula, which also affects the evolution of the star, requires further study. Generally speaking, once the protostellar collapse phase Is over, the star is an object in hydrostatic equilibrium, with a radius a few times that which it will ultimately have on the main sequence, and with internal temperatures of a few million degrees and surface temperatures around 4000 K The star now enters the pre-main-sequence evolutionary phase. The energy it radiates is derived from gravitational contraction, and the initial luminosity is a few times that of the present Sun. The earlier phases of the contraction proceed in the Hert~sprung-Russell (H-R) diagram along the so-called Hayashi track that is, with nearly constant surface temperature and steadily decreasing luminosity. Energy transport in the bunk of the star is by convection during this stage and, for 1 M<3, the time spent on this track is about 10 million yr. The observed "classical" T Tauri stars appear within the first one-third or so of this evolutionary phase. As discussed above, they are observed to be bright in the infrared, suggesting the presence of circumstellar material. This active, classical T Mauri phase is estimated and observed to terminate within a few million years, probably through evolution by accretion or loss of disk material into the naked T Mauri stars. These T Mauri systems appear to be very likely sites for accretion of subplanetaIy and protoplane~ry objects. After the Hayashi phase, the path followed by a contracting star changes direction in He H-R diagram, evolving with gradually increasing
30 luminosity and increasing surface temperature until hydrogen burning ig- nites at the center. During this phase the energy transport within the star is primarily by radiation. When nuclear reactions contribute 100 percent of the energy output, the star is said to have arrived on the "zero age" main sequence. Some aspects of the evolution of the star seem to be reasonably well understood. Others, such as the role of its rotational and magnetic energies in generating and collimating bipolar outflows, still require further and more detailed observational and theoretical study. FORMATION OF THE PIANETS The theory of planetary formation rests on calculations of particle dynamics, collisional accretion theory, and gas dynamics, some aspects of which are still quite uncertain. The committee discusses briefly the concept that all planets formed by essentially the same process, that is, by gradual accretion of small dust particles into larger subplanetary bodies (commonly called planetesimals in this context), which later coalesced to form the planets. This scenano, although not the only possible planetary formation process, is one that is now undergoing intensive study. Even though models based on this picture are oversimplified because of existing limitations of computers, they contribute to the advance of general insight. In these models, the major difference between the inner and outer planets is simply that the latter grew to the point (10 to 20 Where they were able to attract a significant amount of nebular gas to form an envelope around the solid core, which presumably consisted of both rocly and icy material. In the case of Uranus and Neptune the gaseous envelope was much smaller in mass compared to the rest of the planet, and part of the accretion of solid matter could have occurred after dissipation of the nebular gas. In the inner solar system both the heating and tidal ejects of the Sun, as well as the smaller amount of condensable material, apparently prevented the buildup of cores to the Critical mass where significant gas accretion was possible. The starting point is the nebular disk composed of gas mixed with about 1 percent by mass of dust, at temperatures in the range of 100 to 2000 K The dust particles have essentially interstellar characteristics, with typical particle sizes of 10-5 to 10-4 cm (0.1 to 1 ~m). Dust would be absent close to the central star, where temperatures are expected to be high enough to vaporize it. In the inner parts of the nebula beyond this region, out to the point where the temperature falls to roughly 2130 K, the particles are composed principally of compounds of oxygen, magnesium, silicon, and iron. In the outer regions, below 2~ K, water ice can also exist, as well as ices of ammonia, methane, and various clathrates.
31 The first stages of dust accumulation into larger objects are proving to be perhaps the most difficult of all the phases of planetary formation to understand. In one long-standing scenario, dust grains gradually sink to the midplane of the nebula, growing by accretion to centimeter size on a time scale of a few thousand years. Once the dust layer at the midplane becomes dense enough, gravitational instability occurs. The layer fragments into rings, and these further fragment into gravitationally bound aggregates with characteristic sizes of a few kilometers and masses on the order of 10~8 g at the Earth's distance from the Sun. This gravitational instability model of planetesimal formation is simple and appealing, but there are now serious questions concerning many of its basic assumptions and predictions. They focus in particular on the disruptive effects of turbulence in nebular gas, the role of stickiness of dust grains in grain coagulation processes, the physical morphology and settling times of resulting dust aggregates, and the short time scale of ~ 104 yr predicted for accretion of kilometer-size objects (there are estimates that it may have taken 10 to 100 times longer). It is fair to say that at present the mechanisms of growth to this size range, and their required time scales, are poorly understood. There is, however, a widely accepted standard model for subsequent planetary accumulation. Planetesimals, whatever their means of formation, undergo collisions and gradually accumulate into a few large bodies that eventually form the terrestrial planets and the cores of the giant plan- ets. Monte Carlo simulations of the accumulation process suggest that it proceeds rapidly during the first few million years, forming objects up to ~25 percent of the Earth's present mass in the terrestrial planet zone; subsequent collisional growth occurs more slows because Secreting objects decline in number and become orbitally more isolated from the protoplan- ets. Estunates for the total time required to fully accrete the terrestrial planets range from 107 to 108 yr. Once a system of planetesimals foes, it also selves as a reservoir from which dust can be eroded over much longer time scales. As discussed further in Chapter 4, collisional processes could explain the observed maintenance of dust around main-sequence stars 109 yr or more after formation The properties of our own solar system strongly indicate that Jupiter must have formed more rapidly, because (1) the core must have accreted to its present size before the nebula gases disappeared and (2) the presence of the asteroid belt without a major planet in it strongly suggests that the prior presence of Jupiter and its gravitational influence prevented the final stages of accretion from occurring there. Current research is directed toward the question of how to build Jupiter's core, which probably contains about 20 Me, within a few million years (if, in fact, this correctly represents the time scale for dissipation of nebular "gas." There are no observational astronomical constraints on the lengths of time required for disappearance
33 temperatures of a few thousand Kelvin. This phase could possibly be observable. The later evolution involves cooling of the interior and only a very slow decrease in radius; the luminosity and surface temperature decline as a function of time. After 4.5 times 109 yr the present radius and intrinsic luminosity of Jupiter are reached. The above theory of the evolution of the giant planets is based on numerous approximations, does not include rotation and the formation of satellite systems, and will undoubtedly require extensive revision ~ the future; nevertheless, it agrees with many known properties of the giant planets. For example, the theoretically deduced critical core masses are in good agreement with the core masses deduced from observations of the giant planets. The formation process for the outer planets Uranus and Neptune, however, is still not well understood. In particular, the time scale currently inferred for accretion of their cores at their present distances is longer than the lifetune of the nebula and may be longer than the age of the solar system. Perhaps resonances played some role, as suggested by such observations as the trapping of Pluto in a 3:2 resonance with Neptune, and the extraordinary circular of Neptune's orbit. Moreover, it is not understood how these planets could have acquired their gaseous envelopes of approximately 1 Me. Possibly the accretion of Uranus and Neptune may have started closer to the Sun than their present distances. Their cores may have grown sufficiently large to attract some gas from the nebula, and later the protoplanets could have migrated to the outer parts of the nebula under the gravitational influence of Jupiter and Saturn Their buildup could have been completed simply by the accretion of planetesimals. What is certain is that venous critical aspects of planetary formation still need to be clanged. Progress depends on development and testing of improved models for the process, and thus on continuing theoretical and observational investigation of planetary systems. Many theoretical and computational initiatives will require expansion of currently available computer capacity and unproved numerical techniques for solving coupled systems of differential equations. New and detailed observational data on chemical compositions, present physical states, and dynamical behavior both within and outside our own solar system, from spacecraft instruments and from ground and Earth-orbital facilities, are central to this effort.