navigation image map next page table of contents previous page

Models for the Origin of Planetary Systems

This "crash course" on Cosmology culminates with the following brief synopsis on the methods and results of astronomers' search for other planetary systems and consideration of the origin of our own Solar System, plus a quick look at several of the latest, provocative, and especially exciting theories (or brash speculations) on the start of life on Earth and, by extrapolation, the presence of other intellectual beings in our Universe that might be "out there".

There are several JPL movies which you may wish to view before continuing on this page. Access through the JPL Video Site, then the pathway Format-->Video -->Search to bring up the list that includes "Close up View of Plametary Birth", October 8, 2001; "Planet-forming Disks", January 11, 2002; "Beyond the Planets", February 4, 2003; and "Pointing the Way to Exoplanets", December 11, 2003 (this last is a full hour lecture). To start each one, once found, click on the blue RealVideo link.

Two hallmarks in particular distinguish planets from the stars they orbit: First, they usually show marked differences in composition, being either gas balls (whose temperatures are well below fusion levels) with other elements besides hydrogen that have rocky cores or they are dominantly rock with many elements making up usually silicate minerals. Second, they are significantly smaller in diameter (hence volume) than their parent star. This SOHO image of a part of the Sun, with a solar prominence, illustrates this size difference, as displayed by adding a drawn sphere the size of the Earth to allow comparison:

A SOHO image of part of the Sun, with a filled circle representing the size of the Earth.

This huge size difference between the Earth and its normal G star Sun is humbling in its stark truth. The great disparity in size also makes it clear how difficult it will be to find Earth-sized planets around nearby stars - and even more of a technology challenge as astronomers entertain hopes of finding stars elsewhere in the Milky Way galaxy and galaxies beyond.

It is natural for humans to wonder if there is life elsewhere in the Milky Way, and by implication in other galaxies. The starting point in searching for life is to prove the existence of other planets and inventory their characteristics. In the last decade the hunt for planetary systems has intensified. The first extrasolar planet (generally the term "exoplanet" is now in common use) was found in 1995 orbiting the star 51 Pegasi, which lies 50 light years from Earth in the Milky Way (the closest exoplanet found so far is just 15 l.y. away from us). As of June 2003, planetary bodies have now been detected in at least 88 other stellar (~solar) systems; as of September, 2004 a current count of 135 individual planets have been associated with these systems. (Note: Astronomers has disputed some of these observations, so that the firm number of accepted planetary bodies is just under 100.) Analysis of one such system - Upsilon Andromedae -indicates it to have 3 planets (a second triple star system was recently discovered); 8 other stars elsewhere have 2 planets. The U. Andromedae system, diagrammed below, consists of Giant (probably gas-ball) planets (much smaller ones presently are very hard to detect), all within orbits whose distances from its star are comparable to that of the four small terrestrial planets in our Solar System:

The Upsilon Andromedae System of Giant planets around a star in the M.W. galaxy; their relative locations are compared with those of the 4 inner planets of the Solar System.

In August 2004, a star, mu Arae d. just 50 light years away was shown to have three planets, two larger gaseous ones and a third having a size, mass (14 Earth masses), and orbital characteristics (rapid revolution around its parent) that indicates it is likely the first rocky (inner solar system-like) planet yet found. Then, at the end of August, announcement was made of star 55 Cancrie, 41 l.y. away and Cliese 436b, 33 l.y. away having a planet with 18 and 20 earth masses respectively.

So far, only one possible planet (see below) has actually been seen. The others, being too small to be detected optically using present day technology and being low-luminosity bodies, can be deduced to exist from their interactions with their parent star. Almost all discovered so far are large - Jupiter-sized or greater - and are gas balls. Several are extremely close to their star (at distances less than Mercury's orbit; one is thought to be as close as 5 million miles from its star and rotates rapidly around its parent body). Many planets have much more elliptical orbits than those moving in the Solar System. It is hypothesized that most planetary systems will consist of multiple planets but the smaller ones are presently still invisible to us.

A high point in the current search for planets occurred on June, 2002 when several groups of astronomers announced jointly the discovery of up to 20 new planets - including at least two of Jupiter size - in the Milky Way galaxy, whose stellar parents reside at distances ranging from 10.5 to 202 light years from Earth. The closest in has been provisionally named Epsilon Eridani b (its star is the one actually cited in the Star Trek series as the location around which Mr. Spock's home planet, Vulcan, was orbiting). The rate of planet discovery seems to be accelerating. With it is the growing belief by cosmologists that planets could well exist in the billions, i.e., they are the inevitable result of processes that take place when most stars are born. Thus, planets may well turn out to be the norm - the expected, and perhaps the most significant products of nebular collapse in stellar evolution. The proponents of SETI (Search for Extraterrestrial Life; seeking primarily radio signals that have non-random character [perhaps some form of mathematical organization]) have been galvanized by these recent discoveries. (The writer is convinced that it is only a matter of time - probably during the 21st Century - until contact with other intelligent beings is achieved.) We will return to the SETI idea near the bottom of this page.

Current search for extra-solar planets is restricted to the Milky Way and galaxies close enough to Earth for an individual star to be resolved to the extent that changes in its motion can be measured. Gravitation attractions from orbiting planetary bodies cause the central star to wobble. This is the basis for the three methods currently being used to detect anomalies in a star's behavior that lead to the inference of one or more orbiting bodies.

The method that has so far been the most successful in locating (invisible) planetary bodies is called the radial-velocity technique. A component of a star's wobble will potentially lie in the direction on-line to the Earth as an observatory. This to and fro (forward-backward) motion causes slight variations in the apparent velocity of light. That, in turn, gives rise to small but measurable Doppler shifts in the frequencies of light radiation from specific excited elements, as expressed by lateral displacements of their spectral lines. From the wobble magnitude and period, the approximate orbit and mass of its presumed cause - the orbiting planet - can be calculated. This method is sensitive to wobble velocities as low as 2 meters per second. (Jupiter, for example, causes a wobble velocity of 12.5 m/sec imposed on light radial from the Sun. Generally, this method, applied to nearby stars, will detect mainly large planets close to their star but in March of 2000, two planets about Saturn-size (1/3rd the mass of Jupiter) were found.

The second method, astrometry, also relies on wobble but depends on measuring side to side displacements by direct observation through periodic sightings. This determination of relative shifts can be done on photographic plates taken of part of the sky at different times, commonly using the same telescope. But, considerable improvement shift resolution results when two telescopes are positioned apart and joined electronically. This permits application of interferometry such that the two telescopes act as though they were one large one. Resolution as sharp as 20 millionths of an arc second (an arc second is 1/3600ths of an arc degree on the sky hemisphere [0° at the sealevel horizon; 90° at the zenith near Polaris in the northern hemisphere]). The Keck interferometric telescope in Hawaii will soon be operational. This should facilitate detection of even smaller planets in nearby space or large planets in stars 10s of light years away.

As we have seen with the HST images in this Section, resolution (and clarity) is significantly improved by operating a telescope in space, above the distorting atmosphere. The use of two telescopes mounted on a single boom but separated by meters allows the interferometric method to work on space-quality imagery. SIM (Space Interferometry Mission) is a NASA probe slated to fly during or after 2005. This will lead to a millionth of an arcsecond resolution, capable of inferring the presence of planets just larger than Earth-size or of large planets with far out orbits.

A third method is called transit photometry. When a planet's orbit takes it on a path where it passes across the face of the star under observation, the body will block out a small amount of radiation (usually visible light) for the period of transit. Sensitive detectors can note the slight diminution (up to about 2%). To distinguish this from a "transient event" of other origin, the astronomer needs to establish some regularity (reproducible at fixed intervals) of the drop in radiation, which will depend on the nature of the orbit (ellipticity; distance, etc.) Depending on planet size and proximity, the drop in stellar luminosity will be a few percent or less (an accurate determination helps to establish the planet's actual size). Such an effect was first noted in 1999 when a giant planet (earlier found by the radial-velocity method) passed in front of a star (HD 209468) whose light intensity underwent a drop of 1.5%. In a June, 2002 meeting, two other groups using telescopes in Chile have reported 3 and 13 possible transit detections, but these observations have yet to be confirmed independently.

This promising transit approach will be used in the recently approved Kepler mission (launch provisionally set for 2006) in which a space telescope would be pointed on the same areas in the stellar field for up to four years, integrating brightness measurements to single out variations as small as 0.01% (capable of detecting Earth mass-sized planets). Up to 100,000 stars can be conveniently monitored over time and individuals with multiple planets should reveal the relative number of stars that possess planetary systems. Two other non-U.S. missions, COROT and Eddington, are in the planning stage. Thus, as detectors improve and instruments become spaceborne above the atmosphere, sometime in the next 10-20 years Earth-sized planets should become detectable by measuring drops in stellar light curves owing to transits of large planet(s) across their parent star.

A fourth method has just had its first success in June of 2002. Examine this pair of images, shown first as a photo negative plate and then in color:

The <i>winking</i> star method of planet detection

KH 15D observed in color

This is called the eclipsed or "winking" star method. In the left image (a photo negative), a Milky Way star KH 15D (2400 l.y. away; about the size of the Sun) is visible behind a much closer (or larger) star. In the right image it is totally absent, a condition lasting for about 18 days, and then it reappears. This on-off cycle occurs every 48 days. The eclipsing body could not be another star nor is it likely to be a huge (star-size) planet. The interpretation is that there is a cloud of asteroids and dust in a smeared-out clump orbiting KH 15D which block the starlight when a clump passes across the parent star; speculation considers that there may already be one or more planets formed from this debris. There may actually be two clumps (symmetrical pairing) at opposite positions in a single orbit; this has yet to be confirmed. A further anomaly: examination of photographic plates taken many years ago (although at limited intervals) does not detect this on-off phenomenon.

A fifth method is still in the experimental planning stage. The Terrestrial Planet Finder program at Princeton University is developing a special type of "cats-eye" mirror that will greatly reduce the effect of the luminous parent star. When this technology is deemed ready, they hope to persuade NASA or some other agency to use mirrors on several telescope-bearing satellite flying in formation and spaced to utilize the principle of interferometry to improve resolution and to detect small planetary objects.

The ultimate dream is to directly visualize individual planets. This may be possible using several HST type spacecraft flying in formation ("clusters") with separations of a few hundred meters to hundreds of kilometers. In one mode, data will be combined using interferometric principles. Light from the central star can be blocked out by specialized image processing, leaving a residue of low luminosity orbiting bodies detectable by resolution- and radiation-sensitive interferometry. Both NASA and ESA each have in the planning stage such a mission (called The Terrestrial Planet Finder and Darwin, respectively).

However, in September, 2004 a reasonable claim has been made by astronomers using the European Space Agency telescope in Chile of having found the first actual planet. Look at this infrared image of their observation:

The small brown dwarf star 2M1207 (blue-white) and an apparently associated large planet (red).

The star, 2M1207, just 50 light years away, is a brown dwarf (too small to initiate hydrogen fusion). At a distance twice that of Neptune from the Sun is a reddish object five times the size of Jupiter but is cool (less than 2000° C) and has a spectrum that includes heavier elements. This object is likely a planet but final confirmation must await future observations of its changing positions as it orbits (there is a small possibility it is another object beyond the dwarf star).

A stronger case was presented in March, 2005 by a group of astronomers using the European Southern Observatory. They have obtained a picture of GC Lupi with a definite planetary body distanced about 1 Neptune orbit from the star (a; with an extended atmosphere). This planet (b) is about 2.5 times the mass of Jupiter and has a surface temperature between 3 and 4 thousand degrees Celsius. Here is a telescopic view:

A hot planet orbiting star GC Lupi.

Statistically speaking, the number of such planetary systems in the Universe should extend into the millions within individual galaxies and the billions when the whole Universe is considered. It would logically be likely therefore that non- or weakly-self-luminous bodies, i.e. planets, are the norm orbiting around a central star for at least some of the size classes on the Main Sequence of the Hertzsprung-Russell diagram. As such stars proceed through their developmental stages (before they leave the Main Sequence), planets seem the inevitable outcome of the formational processes involved in star-making. So far, however, when the number of stars that have been studied using any of the above detection techniques are used as a base line, only about 5% have yielded evidence of associated planets; this number is probably a minimum for determining the actual percentage that do have planets since smaller ones cannot yet been recognized as present.

Two scientists, C. Lineweaver and D. Grether of the University of New South Wales in Australia, have recently published a study that relies on reasonable probabilities to estimate the number of planets just in the Milky Way. They argue that, of the approximately 300 billion stars they calculate to be the total population of the M.W, about 10% or 30 billion consist of stars similar to our Sun and most likely to have favorable conditions for planet formation. Assuming that, of these, at least 10% will produce giant, Jupiter-like planets; thus their earlier number estimates 3 billion giant planets. Such large planets would almost certainly be accompanied by smaller ones formed out of the materials (they call "space junk") associate with the parent star. These giants help in the collection process that leads to smaller companions. But, more importantly, the giants serve as the principal attractors that gravitationally pull comets and asteroids into them (remember the Shoemaker-Levy event discussed in Section 19) and thus function as "protectors" of the small planets by minimizing the impacts these receive. Now, in a more recent presentation at the 2003 International Astronomy Union in Sydney, Australia they have raised their estimate to perhaps as much as 30 billion giant planets and a similar number of earthllike planets. This bodes well for future hunts as observational technology improves. Although this may seem "wildly optimistic", the likelihood of life on planets (see below) continues to rise dramatically with the increases in estimates of planetary occurrences - especially if one presumes that planets are the norm around stars in size ranges no greater than 10 solar masses.

For a while, astronomers assumed that most stars with planets would be relatively small - Sun-sized to perhaps 10 solar masses. These stars last for billions of years and thus favor the eventuality of life if planets developed during the stellar formation process. Now, several notably larger stars in the Milky Way have been found to have large planets. So, planetary formation is a function of process primarily and may have little to do with how long its star can survive. But the really big stars, even with planets, would burn their fuel and destruct long before evolution would likely foster even primitive organic matter.

The exoplanet systems probably show a wide range of individual types. Our Solar System is but one of many combinations of small-large, rocky-icey-gaseous planets. The results mentioned above are biased towards larger planets and give no real indication of the actual number in their systems. Some systems however may be almost like a binary (ternary) star system in which the second or more planets approach brown dwarf mass but still too small to initiate nuclear burning.

The expectation is that planets have formed over most of the time that stars have developed in galaxies. One star pair (one a pulsar; the other a white dwarf) in a globular cluster within the Milky Way some 5600 l.y. away (in the constellation Scorpius), has been shown to be perhaps as old as 13 billion years. This is based on the sparcity of elements of atomic numbers higher than helium. The large planet (4 times the diameter of Jupiter) now associated with it, is almost certainly the same age since prevailing theory holds planets to form roughly at the same time as parent stars. It verified as a true planet, it must be a hydrogen/helium gas ball similar to Jupiter. It is likely devoid of life owing to the turbulent history reconstructed for this star pairing and to the absence of life-forming elements. Even if some form of primitive life did form on it or associated planets, those would have perished in which harsh conditions prevailed in its later years. But the chief implication of this observation (reported in July 2003) is that planetary formation can be traced to the early days of the Universe and, as carbon accumulated from the many early supernova explosions, some planets may have developed life of some type(s) since the first few billion years of cosmic time or even earlier.

While astronomers exercise caution about conclusions that specify the number of earthlike planets to be expected from their estimation procedure, they do propose that that number should be in the millions. Whether such planets also harbor life is much harder to pin down numerically but their statistical approach suggests that a significant fraction of the earthlike planets would possess the proper conditions. How much of that is intelligent life is still guesswork. The current sample of 1 (us!) is the only data point. But if the reality is actually a much larger number, then, purely from statistical logic, we should expect that some of these intelligent civilizations should be more advanced than ours. Why we have yet to "hear" from them remains uncertain (but now SETI improves the chances for this) unless there is some fundamental reason that makes space travel, even from nearby stars, very difficult.

Now, let's turn to consideration of the ways in which planets form. For planets in general, terrestrial and gas envelope types, dust must be present in sufficient quantities to collect as cores or to comprise the main body of the planet. (As a corollary: if gases dominate, they must remain below temperatures at which fusion can start). There is plenty of dust in galaxies, mixed with gases from which stars emerge. The source of the dust has been somewhat problematic but a prime candidate is exploding stars large enough to synthesize silicon, oxygen and other heavier elements. A recent observation strongly supports this:

A supernova 11000 light years away in the Milky Way.

This 300 year old supernova is hard to see in visible light because of a superabundance of dust. When imaged in submillimeter light, the above pattern stands out. The brightest areas are advancing gases with large quantities of dust that give off light at these wavelengths.

One essential requirement for a planetary system to develop is that it forms during the organization of a central star (possible exception: a captured planet, probably rare). Also critical is the availability of dust and gas. The processes involved can be somewhat varied but are sensitive to a relatively narrow range of conditions. The sequence of formative events probably begins during the T Tauri stage of developing stars in which the conditions are favorable. These stars have notable dust clouds (nebulae) that can be monitored in the infrared. Some evidence indicates the clouds will begin to reorganize their tiny particles into large clots, which can grow to planet size, in about 3 million years. But, detection of cold nebular material at longer wavelengths suggests the dust can take 10 million years or more to build up any planets that may result.

Another important factor, recently reported, is that stars which have a relatively high content of iron (from gases enriched by repeated mixing of supernova explosions over time) will have a much greater likelihood of producing planets. Iron is a measure of the metallicity of a star (page 20-7). The iron may be needed to develop planetary nuclei (most ending in planet cores) that help in the gravitation attraction that drives accretion through collisions and infall. Stars with three times the iron content of the Sun have an estimated chance of having planets set at 20% (This comes from a study of 754 nearby stars in the current (on-going) inventory of which 61 have detected planets (this amount to a probability of about 8% for all stars of mass less than 10 times the Sun having auxiliary planetary bodies). The results of this study are depicted in this graphical diagram:

Metallicity diagram of Fe distribution in 754 nearby stars.

The paradigm summarizing the processes involved in the formation of the Sun and its planets probably applies (with variations) to most other planetary systems in general. The first realistic notion of how planets form was proposed by Pierre Laplace in the 18th Century. In its modern version, both stars and their planets are considered to evolve from individual clots or densifications within larger nebular (cloudlike) concentrations dominantly of molecular hydrogen mixed with some silicate dust particles that spread throughout the protogalaxies and persisted even as these galaxies matured. In younger stars, much of the hydrogen and the heavier elements are derived from nova/supernova explosions that have dispersed them as interstellar matter that then may initiate clouds or mix with earlier clouds. Such nebulae are rather common throughout the Universe, as is continually being confirmed by new observations with the Hubble Space Telescope.

One of the best studied and, in itself spectacular, is the Orion Nebula, seen here:

HST view of part of the Orion galaxy, showing billowy nebular matter in process of the clumping of material that will generate new stars.

Below are three views of nebular materials associated with the famed Eagle Nebula: Top = full display of the M16 (Eagle) nebula (note the dark dust areas; the white dots are stars lying outside this nebula); Center = part of Eagle nebula, showing top of the Horsehead pillar of dust and gases from which stars and planets may eventually evolve (a few stars already being evident), made by combining three film exposures through the Kuevn Telescope at the European Southern Observatory ; Bottom = details of the temperature variations in the dust making up the solid particulates in the Eagle nebula as imaged by the European Space Agency's (ESA) Infrared Space Observatory (ISO) at two thermal infrared wavelengths, in which red is hot and blue is cooler (about - 100° C):

M16 - the Eagle nebula

Detail of the top of the Horsehead Pillar, in the M16 Eagle Nebula.

ISO thermal image of part of the Eagle nebula; blue made from 7.7 µm exposure; red from 14.5 µm band.

As individual stars start to develop within these gas and dust clouds, in many instances that dust will organize into a protoplanetary disk (see third figure below). The NICMOS infrared camera on the Hubble Space Telescope has observed a prime example of this stage, in which the glowing gases moving into the central region where a protostar is building up are cut by a band of light-absorbing dust that is most likely disk shaped (can't be verified from the side view in this image):

Gases heated to incandescence that are presumably collecting into a central star; the dark band cutting across the bright field is almost certainly a protoplanetary disk from which planets may emerge as the star becomes fully organized.

An Earth-based telescope has captured this view of a nearby star (nicknamed the "flying Saucer"), again with a girdling disk of dust (and an as yet unexplained anomaly in that the upper half of the image is redder than the bluer lower half):

Disk of dust around a star, in a photograph made through the European Space Observatory's New Technology Telescope.

Examination of these gas and dust clouds by HST has led to the discovery of small clumps or knots of organized gas-dust enrichment within the protoplanetary disks called Propylids found in the neighborhood of a parent star. This may be a more advanced stage of concentration that results in a new star with an envelope of gas-dust suited to accretion that produces planets. Three such propylids are evident in this image of the nebula associated with the Orion galactic cluster (go back to the top image in the three above to try to spot for a conspicuous propylid).

Propylids in Orion.

Other individuals, at least one possibly as recently formed as 100,000 years ago, found during the Orion study (see page 20-2 for a view of the entire young nebula in the Orion group) look like this in closer views:

Additional protoplanetary disks in Orion.

Propylids are vulnerable to being destroyed by UV radiation from massive, young nearby stars. It is surprising therefore that many propylids (some shown below) have survived in the Carina Nebula, which has numerous UV-emitting stars. Other factors must be involved

A sampling of propylids in the Carina Nebula.

Development of planetary cores, from which full larger planets then form, is a race against time. Examine this model:

Model of protoplanetary disc evolution (credit Ann Field)

The major threat to protoplanet formation is the stellar (solar) wind coming from the parent star (Sun). UV radiation is also able to break apart the smaller particles. Wind and UV radiation are capable of pulling apart small particles. These particles in the early stages are tiny bits of dust (solid; solids with an ice coating; ice) which are charged electrostatically. These will collide from time to time and may stick. If not disrupted by the stellar wind the now larger particles can again bump with others. By the stage in which some particle conglomerations have reached baseball size, they can resist stellar wind forces and survive to grow ever larger through collisions.

Thus, as the gas and dust cloud forms around a growing star, particles of solids begin to clump and some survive disrupting actions. However, much of the gas and dust may be pushed continuously away by the wind and radiation, so that the amount of material available to form planets generally diminishes over time. Evidence suggests that most propylids are blown away before a planet grows large enough to survive, implying that the planet formation process may be less efficient and common than had been thought during the last decade. If the planetary cores do build up fast enough, they will survive the expulsion of the bulk of the gas/dust. This phase of planet formation occurs typically in a time frame of just 100,000 years or so; it is estimated that 90% of such clouds are dissipated before significant planetary cores can form. Planet accretion leading to survival is estimated to take up to 10,000,000 years.

Most galaxies began during the first half of the Universe and contained a large number of massive stars that formed early in galactic histories. These galaxies have continuously been evolving through the eons as supernovae synthesized elements (see page 20-7) and dispersed them, and new, mostly Main Sequence stars, chemically enriched with elements of higher atomic number, have continued to form well into younger times up through the present. The smaller stars have lasted much longer and are probably the preferred sizes suited to planetary formation and survival. Even today stars are developing from the gases contained in remaining nebular material, so new planets could still be forming. What is not yet known is the percentage of stars in a galaxy that actually have planetary systems. The low number so far found does not necessarily represent an indication of sparsity, since planets are so small and most low in luminosity (mainly from reflected light) relative to their parent star that present direct observations will produce extremely low numbers, thus subjecting our current conclusions that planets may be rare to a misleading bias. Some models of star formation from gaseous nebula suggest that a fraction of the gases, dust, and free molecules is trapped in orbit without infalling to the central star and can organize in planets of various sizes, distances, and composition in a manner similar to our Solar System. If a valid argument, this says that planetary bodies are commonplace throughout the Universe.

A telescope observation, reported in April, 1998, records the sighting (through the Keck II telescope on Mauna Kea, Hawaii) of what is interpreted to be another "solar" system around star HR 4796 (about 220 light years away). This image (at a resolution in which individual pixels stand out), taken in the IR, shows this central star (yellow white) surrounded by a lenticular (in an oblique view), flattened disk of gases and solid matter (glowing hot [reds] in the infrared):

Keck II IR image of a possible solar system around star HR4796.

The diameter of the lens is about 200 A.U. No evidence of individual planets can be made out but the discoverers judge this feature, which has caused quite a stir of interest, to be an emerging planetary system in a "young adult" stage of development. It will certainly be a target for more detailed HST observations.

More recently, the Hubble Space Telescope has imaged star HR4796A, in our galaxy, which shows both a disk and irregular dust and gas clouds. This disk is interpreted to be in a more advanced (mature) stage of development than the protodisks shown above. Although not discernible, there may be planets already in the evolving gas/dust cloud which is made visible by the star's light.

Illuminated dust and gas in the disk surrounding state HR4796A; HST image.

A very well-developed debris disk occurs around the young (200 million years) star Fomalhaut, the 17th brightest star in the sky (southern hemisphere), just 25 light years away. The disk, about 150 A.U (20 billion kilometers) from its off-center star, is analogous to the Kuiper Belt of debris around our Sun. It appears to consist of icy dust. Time-motion studies indicates that there may already be one or more planets that are influencing the large objects in the disk, much like small moons perturb Saturn's rings. Three views of the Fomalhaut disk, the first with the Spitzer Space Telescope, the second with IRAS, and the third, made from HST imagery, appear here:

The Fomahaut star system, 25 light years distant, with different images made at different wavelengthsl; in the 2470 micron image the central yellow circle may be due to zodaical light equivalent.

Variations in density of matter in the Fomalhaut disk.

The debris ring around the nearby star Fomalhaut as revealed by special processing of HST data, which indicates a strong concentration of icy dust in the outer part of the material cloud.

The HST has also imaged two other distinctive debris disks around a central star, as shown in the next illustration. The disk on the left is caused by reflected star light off myriads of small particles; the star, blocked out, is a red dwarf. On the right is another disk seen at nearly normal to a blocked out red dwarf 88 l.y. from Earth that here appears orange because of a different combination of spectral bands used to create the image. This star-disk system may be as young as 250 million years since its protostar began to burn.

Twp HST images showing planetesimal debris disks around the stars (labeled, but blacked out in this rendition).

Theory indicates that, in the earlier stages of planetary formation, some number of broad rings should develop at various distances around the central star. One or more of these would appear as torus like glowing collections of dust and gases. At least two stars with this feature were imaged by HST and reported in January of 1999. Here is one of the star-ring systems:

 Ring of nebular material around Star H141516, which some claim is similar to conditions that could lead to planet formation; HST image.

Visible is a bright ring at some considerable distance out from the tiny parent star (white dot) and a more diffuse, darker mass extending beyond, both features occupying a flattened disc. In this instance, there are no rings close in (analogous to the regions occupied by the inner planets of the Solar System). The white circle is added by the astronomers to mark the boundary within which no visible planetary disk matter has been detected; the broad black cross (X) is an optical artifact. This star is about 350 light years from Earth.

A recent image, made from data detected at 1.3 mm by a French radio telescope, may have caught the formation of two large clots of matter likely to eventually contract into giant planets. These occur in the ring around the central star Vega, 25 light years away (in the Constellation Lyra). Here is an image based on observations made by D. Wilner and D. Aguilar of Harvard's Smithsonian Center for Astrophysics. (Note: this image has been enhanced artificially as an artist's rendition.)

The ring of gas and dust around the star Vega, with the yellow clots being possible sites of planetary formation.

C. Chen and M. Jura, Univ. of California-Berkeley have detected (monitoring infrared radiation through the Keck telescope on Mauna Kea) a ring or disk of dust which seems to contain asteroid-like bodies around zeta Leporis, a star some 70 l.y. from Earth that is twice the mass of Sol and 15 times its luminosity. The ring, first detected by IRAS, is much closer (~5-12 A.U.) to its parent star than the distances found for other recently observed or inferred planetary bodies around stars. Although imagery of the star, HR 1998, does not reveal directly these bodies, their presence is inferred from their average temperature of 350°K (77°C or 170°F). From the data they accumulated. Chen and Jura have produced a plot of the asteroidal ring around zeta Leporis, with a comparison of the Sun's asteroidal belt also displayed:

Diagram showing the general location of a broad asteroidal ring around zeta Leporis, compared with the belt around Sol, our Sun.

Chen and Jura propose this ring to be the precursor of eventual formation, by collision of asteroidal bodies, of rocky planets analogous to those of the Solar System. These bodies, form from smaller particles (dust) condensed from the gas-particle cloud associated with the forming star. Much of that dust can move inward towards the star by a process called Poynting-Robertson drag. This is caused by radiation from the parent star being absorbed and re-radiated differentially, leading to a Doppler effect (here, the energy of emission in the direction of dust motion is at shorter wavelengths [more energetic] and thus by retro-action slows the particles) that promotes drift of the dust towards the star.

From the above, one necessary step in planetary formation is the development of a protoplanetary disk around a suitable star (or a binary star pair). An inventory of detectable disks in nearby neighborhoods of the Milky Way found disks around 236 stars; the instrument used was the Spitzer Space Telescope whose infrared sensons are adept at measuring hot gas and particles. As evident in the graph below, most of these disks are associated with younger stars but the distribution includes even older (>800 million years) stars:

Distribution of gas-dust disks surrounding nearby stars.

The following is a generally accepted model (called "core accretion") for establishing a planetary system: A nebula is subject to gravitational irregularities and other perturbations that cause free-fall collapse to numerous clots around which surrounding gases and particulates usually adopt a disklike form. Over time, the disk tends to organize in spiral arms of gas and matter, which increasingly become disorganized by clotting (discussed below). Consider this generalized sequence:

Simplified model of star-planet formation involving a disk which evolves spiral arms of clotting matter until this structure is broken up (lower right) as individual planets begin to form.

Two more extrasolar examples of these planet-star formation gas/dust disks appear below; read the caption for details.

Disk around star HD141569; density variations are colored in red-yellow tones. A companion star, without much dust, is in the upper left. In the right view, the image on the left, seen at an angle, has been rotated by computer processing to appear as if the astronomer is looking straight down at it.

The Epsilon Eridani system, 10 light years away, imaged in submillimeter wavelength light. Yellow clumps may be associated with planets.

The influence of gravity, which builds up progressively as planetesimal clots enlarge, is the prime driving force promoting both planet and star (Sun) formation. In some instances, shock waves from a supernova can cause interstellar matter to initiate collapse and compress into protostars and debris orbiting them. Matter is also redistributed along magnetic field lines by magnetohydrodyamic processes. The main phases of planetary formation extend over about 10-20 x 106 years but it may require up to 108 years to progress from the early infall to the late T Tauri stage of a protostar's development. While a particular clot is organizing, the materials tend to redistribute such that hydrogen and much of the lighter elements flow towards a growing center to accumulate in a gravitationally balanced sphere, the star. Under one set of conditions, instabilities lead to a double (binary) star pair. As protostars form, the rotating gases and dust particles collect in a spinning disk around each center and eventually organize by accretion into planets. The same process, with variants, works at single stars. The time frame for the above model suggests planets to appear within one hundred million years or less after the nebular gas and dust have begun to behave as a unit in space.

If our Solar System is the norm, inner planets should be rocky, with thin or absent atmosphere (lost from insufficient gravitational ability to retain the gases or by being swept away by the solar wind). Outer planets should have rocky cores and be less susceptible to loss of gases, so that their increased mass serves to gather in still more gases. However, the discovery that giant planets can lie quite close to their parent stars places this assumption of size distribution with distance into question.

Alan Boss of the Carnegie Institution of Washington has argued that the outer gas planets Uranus and Neptune have much less gas that would be expected from conventional planetary system models. He claims that karge quantities of gas were driven away in the earlier history of these planets by UV radiation from nearby stars in the local cluster. It is reasonable to expect that stellar windw, UV radiation, and other "forces" from neighboring stars might affect planetary history but his hypothesis remains in dispute.

Two recent hypotheses are adding new twists to the above concepts. First, in addition to or in place of core accretion, another mechanism called "disc instability" may play an important, perhaps key role, in planetary inception. This is related to gravitational irregularities that can cause rather rapid accumulation of materials in the proto-planetary disc. Earlier-formed planets can contribute to setting up further instabilities. A second idea holds that planets can move inward or even outward in a form of migration or "wandering" so that their orbits change both in relative distance to their parent star(s) and in eccentricity.

But for the present, astronomers continue to build and refine their models on the much easier-to-make observations at the astronomically short distances within the Solar System. Like other stars, the Sun (whose diameter is 1,392,000 km [870,000 miles]) is an end-product of gravitationally-driven condensation and collapse of hydrogen/helium gases and associated matter (both solid and gaseous) consisting of other elements and compounds that once made up a diffuse (density ~ 1000 atoms/cc) nebula. Probably many stars were generated in the timeframe of a few hundred million years from this particular "cloud".

The protosun built up from centripetal, gravity-induced infall of nebular substances towards one of the concentration centers in the nebula. The bulk of the gases enters the resulting star itself along with much of the other materials, leaving an enveloping residue of matter enriched in Si, C, O (and H), N, Ca, Mg, Fe, Ti, Al, Na, K, and S (most organized into compounds, particularly silicates, that can be sampled by recovery of iron and stony meteorites - representing fragments of comets and broken protoplanets that are swept up onto Earth). This material, bound by gravity to the Sun but free to move inertially in encircling orbits, remained distributed in the space making up the Solar System. This system of particles rather rapidly organized into a disc-like shape whose present radius is about 100 A.U. (Astronomical Unit, defined as the average distance [149.6 x 106 km, or about 93 million miles] between the centers of the Earth and Sun; solar light takes about 8 and a half minutes to travel that distance; Pluto, lies 39.5 A.U. from the Sun whose gravitational influence is exerted well beyond). The disc rotated slowly (counterclockwise relative to a viewpoint above the north celestial pole [which passes through Polaris, the North Star]), its motion influenced by external gravitational effects from nearby stars.

As this rotation got underway, and thereafter, the stellar (solar) magnetic field churned up the dust and gases (descriptively compared to the action of an "eggbeater" in a thin batter) causing them to collect into clots much smaller than the Sun that underwent various degrees of condensation. This field also expels and guides this material into jets that carry matter out to great distances, as seen here in this Hubble Space Telescope view of a jet ejecting from another star in our galaxy:

Hubble image of a jet ejecting from another star in our Solar System.

Both jets and irregular nebular patches (e.g., the Horsehead and Eagle nebulas shown above) contain not only gases but significant amounts of dust. The dust is very small and consists of three types: 1) core-mantle elliptical particles, typically 0.3 to 0.5 microns in long dimension, with a silicate interior coated by icy forms of gases; 2) carbonaceous particles (~0.005 microns), and 3) open frothy clots called PAH dust (polycyclic aromatic hydrocarbons) (~0.002 microns). Shock waves and radiation can strip off the ice mantle leaving grains that are incorporated into coalescing bodies that form the prototypes which accrete into the planetesimals from which asteroids or planets then build up. Ultraviolet radiation can modify the organics into more complex molecular forms. In this way, organic molecules are introduced from space onto planetary surface and, if conditions are right, can eventually serve as viable ingredients for the inception of living things (see below).

The possible role of shock waves in planetary formation is now the subject of considerable study. Evidence for a shock wave that develops as material falls towards a nearby protostar against its remaining gas/dust cloud has been observed at L1157, in which the present cloud is about 20 times the solar system diameter. As this cloud proceeds to infall into the newborn star as it organizes into a disk, it produces shock waves that may clump dust together, as described in the next paragraph. Here is this cloud:

Nebular dust envelope around a protostar in L1157.

For the Solar System, shock waves and intense radiation acted on the dust such that some of it melted into tiny droplets which chill into chondrules. These spherical bodies then were caught up with remaining dust to produce the primitive small solid bodies (fluffy "rockballs") that populated much of the heliosphere surrounding the Sun. We can analyze samples of these accreted bodies today as meteorites which are small pieces of them torn loose and put in orbits that eventually reach Earth. (You can review some basic knowledge of meteorites by clicking to page 19-2.) Most infallen meteorites are ordinary chondrites that, in thin section, appear much like this sample from the Tieschitz meteorite:

 Photomicrograph of a thin section through the Tieschitz chondritic meteorite in which the round objects are crystalline chondrules.

The most primitive meteorites, called carbonaceous chondrites, are enriched in carbon and contain water. Other meteorites are iron-rich (some with > 90% metallic iron), and may have once been the interiors of planetary bodies since disrupted. The chondrules themselves generally show a very limited size range, suggesting that ones larger than these fell back into the Sun through gravitational pull whereas smaller ones were swept away into interstellar space through expulsion by shock waves and solar wind.

Magnetically-driven eddies within the gas/dust cloud helped to impart additional angular momentum to the larger condensed rotating objects beyond the spherical Sun (which possesses only 0.55% of this momentum even though it contains 99.87% of the total mass of the system). These objects now remain in orbits around the Sun in positions that remain stable because of the counterbalance between centrifugal forces related to angular momentum and inward-directed gravitational pull from the Sun.

Planets appear to form simultaneously with the star around which they associate in well-defined orbits. Two general models for planetary formation (mainly of large planets with thick gas atmospheres) - Accretion and Gas Collapse - are popular now, and both may have operated. These models are shown in these two panel sequences:

Two models for planetary formation; applies primarily to Giant gas planets.

For the Accretion model, as the formative process operated, local instabilities in the nebula tied to the Sun caused the chondrule-laden rockballs within turbulent zones to cluster and further aggregate into objects ranging from meter-size up to planetesimal dimensions (tens to a few hundred kilometers, typified [perhaps coincidentally] by asteroid proportions).

Planet formation diagram.

From J. Silk, The Big Bang, 2nd Ed., © 1989. Reproduced by permission of W.H. Freeman Co., New York

During this growth stage, smaller planetesimals tended to break apart repeatedly from mutual collisions while larger ones survived by attracting most of the smaller ones gravitationally, growing by accretion as new matter impacted on their surfaces. Once started, "runaway" growth ensues so that many planetesimals combine into bodies that eventually enlarge into fullblown planets. The bulk of the matter beyond the Sun was swept into the planets and their satellites, although some remains in comets and cosmic dust. Mercury, and some Outer Planet satellites are preserved remnants of this later stage in planetary growth, as indicated by their heavily cratered surfaces that were never destroyed by subsequent processes such as erosion. In contrast, the Moon appears to have built up by re-aggregation of debris hurled into space as ejecta from a giant impact on Earth soon after our planet formed; once collected into a sphere (which probably melted), the lunar surface continued to be bombarded with its own remnants as well as asteroids and other space debris. Its oldest craters are hundred of millions of years younger than the time at which the debris reassembled, melted and formed the lunar sphere; at least some of its larger basins are somewhat older.

The formation of the Moon by collision between two separate but large planetary bodies likely was not a unique event in Solar System history. During the earlier stages of accretion more planets than now exist probably formed. Their number was reduced by a collision which could have caused both bodies to be disrupted into particles of varying size in a "collision cloud" which then reassembled into a single planet whose mass was approximately that of the two earlier bodies (some matter was lost to space). The asteroid belt between Mars and Jupiter may be an example of collisional debris that failed to reorganize into what would have been the fifth rocky planet.

Artist's illustration of one scenario involving a two planet collision.

As mentioned above, in our Solar System, the four inner planets (the Terrestrial Group) are largely rocky (silicates, oxides, and some free iron; three with atmospheres) and the outer four (Giant Group) are mostly gases with possible rock cores. These Giants developed large enough cores to attract and capture significant fractions of the nebular gases dispersed in the accretion disk.

Analysis of argon, nitrogen, and other gases in Jupiter indicates their amounts are such that this Giant must have formed under very cold conditions; if further work bears this out, Solar System scientists may adopt, as one plausible explanation, an origin of Jupiter (and perhaps the other Giants) at much greater initial distances from the Sun with these having since moved significantly closer through orbital contraction or decay. The ninth planet, Pluto, the smallest and, at times, farthest out (its elliptical orbit periodically brings it within that of Neptune), appears to be made up of rock and ice and may be a captured satellite of Neptune.

Theoreticians differ as to the exact methods and sequence in which the planets accumulated after the condensation and planetesimal phases. Timing is a critical aspect of the formation history. One version - the equilibrium condensation model - considers condensation to happen early and quickly, in a few million years, with the observed sunward zoning of higher temperature minerals and greater densities in the rocky inner planets both being consequences of the increasing temperature profile inward across the solar nebula. Accretion was stretched out over 100 million years or so. The heterogeneous accretion model holds condensation and buildup of planetesimals to proceed simultaneously over a few tens of millions of years. Neither model adequately explains the fact that both high and low temperature minerals aggregate together in the inner planets to provide materials capable of generating the atmospheric gases released from these planets. The models also do not fully account for the strong preferential concentration of iron and other siderophile ("iron-loving") elements in the inner, terrestrial planets. One solution is to add (by impact) low temperature material to the growing protoplanets carried in along eccentric orbits from asteroidal and giant planet regions. This material is then homogenized during the total melting assumed for each inner planet early in its evolution (this melting is the consequence of heat deposited from accretionary impacts, from gravitational contraction, and from release during radioactive decay). As cooling ensues, materials are redistributed during the general differentiation that carries heavy metals and compounds towards the center and allows light materials to "float" upwards towards the surface.

Much less is known about the evolutionary history of planets and their eventual demise (destruction). Extrapolating from our Solar System with its two major types of planets - Rocky and Gaseous - and the variety of surfaces on them and their satellites, it is evident that a great range of sizes, compositions, and surficial states can be expected among the millions of planets that many believe exist in the Universe. In the Solar System its complement of planets have survived essentially intact (possible exception: the asteroid belt) since the Sun itself organized some 4.5 to 5 billion years ago). The Sun is expected to burn out its fuel in another 5 billion years, the expanding rapidly into a Red Giant. The outward surge of its gaseous envelope should consume many - maybe all - of the named planets as well as other solar material. This is probably the usual mechanism of most planetary destruction - consumption by Red Giant expansion or by novae or supernovae (see top of page 20-6). Another possibility: gravitational pull brings the planets into their parent stars. Generally, planetary systems around massive stars, if indeed these do form, will be short-lived as those stars themselves do not last billions of years (thus, such stars are not likely to harbor life since not enough time elapses to permit development by evolution [see below]). Smaller stars, such as G types, are much more favorable bodies for fostering life on any planets that may revolve about them, owing to their longer spans of existence.

After the above was written, the writer was made aware of a second form of dust around stars that has been produced after they were formed and well along their trail of evolution. This appear in the April, 2004 edition of Scientific American in an article by Davod Ardila entitled "The Hidden Members of Planetary Systems". (He points out that not all stars with this kind of dust necessarily have associated planets). The dust is of two types: micron-sized particles that are analogous to the dust in the solar inner planet belt that gives rise to Zodaical light at sunsent; dust of a range of sizes that exists further out (beyond Jupiter and the Kuiper belt for our system but for some stars the disk extends out to notably greater distances). The zodaical dust is produced by release from comets and grinding of asteroidal-sized bodies that collide and abrade over time. The larger particles tend to spiral into the parent star, the smaller are pushed away from the star by radiation pressure. Over time, the amount of dust will diminish. But some of the dust may be incorporated in planets within this circumstellar debris cloud, already formed or yet to form. The temperatures associated with dust belts and clouds varies, so that telescope sensors will pick up measurable EM radiation at different wavelengths. One of the best examples of a huge dust disk is found around the star Beta Pictoris, 63 light years from Earth. The disc extends out about 1100 A.U. (about a 460 billion kilometers in diameter). There is a suggestion of one or two planets within the disk. The visible light HST image below shows the symmetrical disk (lower image is colored to indicate density differences); the black center is due to screening out the star itself using a coronagraph accessory on the telescope.

The Beta Pictoris dust disk.

Recently, another prime model for the origin of planets has been reported by Dr. Jeff Hester and colleagues at Arizona State University. The figure below is relevant:

A scenario for star formation in the Trifid nebula in which a wide range of sizes are produced including those of Sun-size.

As with the above models, clouds of hydrogen gas and silicate dust are needed. Shown here is the Trifid nebula. Within it are now being formed a range of embryonic stars which include besides those of Earth-size, more massive stars that explode. The astronomers studying this "nursery" of stars point out that massive stars can produce the isotope of iron Fe60 whose half life is about 1.5 million years. Its stable daughter product Ni60 has been found in meteorites, whose parent sources presumably formed along with the stars like the Sun. This implies that the Sun was born out of a gas-debris cloud that had been enriched by radioactive Fe60 from one or more exploding (supernovae) stars in its neighborhood. These stars were much more massive than the Sun. The nebulae like Trifid, e.g., Eagle and Orion which we've looked at earlier in this Section, are enriched in HII (doubly ionized hydrogen). As shock waves produce ionization of the hydrogen, YSOs (Young Stellar Objects) will form at various sizes. In the ongoing process of star formation, EGGs (Evaporating Gaseous Globules) develop and evolve into associated propylids. The propylids later shed some of their material leaving stars on the Main Sequence of sizes similar to the Sun.

Little has been said about life on planets on this page - this is now reviewed on the next page, 20-12. However, a recent article by Beer, King, Livio, and Pringle in the Monthly Notices of the Royal Astronomical Society has put forth an argument that life must be rare. This is deduced from the observation of the 120+ planetary systems so far discovered. Nearly all of these apparently have only Giant gas ball planets that are much closer to their central star that Jupiter through Neptune in our Solar System. If this turns out to be the case then rocky planets are scarce - those closer to their stars are prone to having much of their gas envelopes blown away by stellar wind and high temperatures proximate to the star. The flaw (and saving grace, for those who want life discovered elsewhere) in the argument is that none of the current methods of planet detection are capable of finding smaller planets but are biased towards locating big gas balls.

There is a growing consensus that planetary systems similar to the planet types and distribution characteristic of our solar system are uncommon, and possibly rare (although most scientists doubt that our system is unique in having life). This estimate is, admittedly, based on the observations to date that show mostly giant planets around other stars. But, while most planets are believed to have rocky cores, stellar wind and explosion processes tend to blow off gases from smaller inner planets. The essential conditions needed for organic molecules to develop are water (preferably in liquid form, but life in steam or ice is believed possible), an appropriate temperature range, some semblance of a favorable atmosphere (but anaerobic or oxygen-free environments on Earth can contain life), and the appropriate ingredients (C, H, O, N, and P; an Si life system, instead of C, is theoretically possible). We shall see on page 20-12a that the Drake equation provides a mathematical means of estimating the opportunity for organic molecules to form on planets in some fraction of the star systems. Likewise, the likelihood for life to occur on planets seems to follow the Goldilocks dicturm (page 20-11a): "not too hot, not too cold, just right".

Having postulated that planets are probably rather commonplace in the Universe, let us study on the next page the types of and conditions for life having formed on Earth by processes becoming ever better understood.

navigation image map next page previous page

Primary Author: Nicholas M. Short, Sr. email: