Printer Friendly

The diversity of planetary systems: new worlds.

Modern astronomy reveals to us, for the first time in history, scenes from one end of the cosmos to the other. We have picturesque views of planetary surfaces in our own solar-system backyard and panoramas of adolescent deep-field galaxies swarming near the limit of the observable universe. But beyond providing pretty pictures, astronomy places our world and our brief human lives in their true contexts - as vanishingly tiny subplots in a truly enormous cosmic play. The curtain opens with a Big-Bang synthesis of the chemical elements that eventually lead to self-replicating, competitive structures of molecules we call "life." While we humans play out our brief bit parts, we yearn to grasp the overall plot.

The scene on Earth today is one of irrepressible biological activity, as it has been for several billion years. So we naturally wonder about life on other planets and moons. No other place in our solar system is teeming with life. But what about worlds beyond our solar system? Are they numerous, and how many of them have conditions ripe for biology?

In the fourth century B.C., Epicurus spoke boldly of the infinite worlds that logically followed from the infinite number of "atoms" that he postulated. His contemporary, Aristotle, differed, seeing Earth as the unique center of a perfect crystalline sky. Aristotle's Earth-centered cosmos dominated Western thought for more than 1,500 years. The notion of other worlds took hold again only after Copernicus yanked the Earth from its central position and placed it in orbit around the Sun with other planets. Soon various thinkers realized that the stars might be distant suns and therefore might have planets of their own.

But for centuries, detecting "extrasolar" planets - those orbiting other stars - seemed beyond all possibility. Shining by reflected light, planets should be roughly a billion times (perhaps 22 to 25 magnitudes) fainter than their host stars. And they would appear separated by less than a few arcseconds, at best, from even the nearest stars in our stellar neighborhood. If extrasolar planets exist, they are lost in the glare surrounding a star's image.

Wobbling Toward Planet Detection

To overcome this glaring problem several research groups are pursuing an indirect method of detecting extrasolar planets. This method has paid off spectacularly in the last three years. Rather than looking for planets directly, it makes use of Newton's Second Law: "for every action, there is an equal and opposite reaction." Just as a leashed dog can jerk its heavier owner around in circles, a gravitationally bound planet will swing its star around in a small mirror image of its own orbit, as they both move around their shared center of mass. Such a stellar wobble betrays the existence of an unseen orbiting body. The size of the wobble tells the planer's mass. The time the star takes to complete one wobble is the planer's orbital period.

The challenge rests in detecting that tiny stellar movement. Our Sun's wobble, caused primarily by massive Jupiter, is a near-circle about 1/1,000 as large as Jupiter's orbit (because Jupiter has about 1/1,000 the Sun's mass). Viewed from 30 light-years away, the Sun would appear to travel around a tiny circle with a diameter of only 1/1,000 arcsecond (1 milliarcsecond) every 12 years. That's as small as a hula hoop on the surface of the Moon would appear from Earth.

The stellar wobbles are so tiny that they are best detected by the Doppler effect they impose on a star's light. As a star approaches the observer, its light waves are very slightly compressed, shortening the wavelengths toward bluer colors. Conversely, as the star recedes from Earth, the wavelengths are slightly lengthened, or redshifted.

This Doppler shift is excruciatingly tiny. The Sun wobbles by about 12.5 meters per second (28 miles per hour). A clear, reliable detection of a Jupiter-like wobble thus requires measurement precision of about 3 meters per second, or easy jogging speed. This causes the wavelengths of starlight to change by a mere 1 part in 100 million. Until recently such precision was far beyond reach, but it is now being achieved by our research group, and three other research groups are not far behind.

The technique we developed to measure this tiny Doppler shift involves placing a cylindrical glass cell of gaseous iodine in the beam of light from the telescope. When starlight passes through the iodine vapor, particular narrow wavelengths of light are absorbed. These dark iodine "absorption lines" serve as tick marks against which the absorption lines in the star's own light can be measured. The measurements are done by sending the starlight into an advanced spectrometer. We are fortunate to use two of the world's finest, both built by Steven S. Vost of Lick Observatory. One spectrometer is currently located at the Lick Observatory 3-meter telescope, and the other is at the 10-meter Keck I telescope in Hawaii.

The spectrometer spreads the iodine-marked starlight into its composite rainbow of colors, or wavelengths, focusing the final spectrum onto a CCD detector at extremely high resolution. We determine the Doppler shift by synthesizing, on a computer, the star's theoretical spectrum for different "trial" values of the Doppler shift. Invariably, one of the trial Doppler shifts yields a synthetic spectrum that fits the actual observed spectrum to an accuracy limited only by the noise that remains in the data. This seemingly simple process requires massive computer time.

Once we find it, the best-fit Doppler shift includes not only the star's motion with respect to the Sun but also the telescope's own motion on an Earth that's rotating around its axis and revolving around the Sun. We use a very precise ephemeris of planetary positions from the Jet Propulsion Laboratory to remove all motions of the Earth, even tiny gravitational perturbations by other planets of the solar system. This leaves us with the target star's net Doppler shift. We currently measure radial (line-of-sight) velocities of stars to within 3 meters per second, adequate to detect alien Jupiters.

Since 1995 eight extrasolar planets of sunlike stars have been positively detected by the Doppler technique. The first was discovered orbiting 51 Pegasi by [TABULAR DATA OMITTED] Michel Mayor and Didier Queloz (Geneva Observatory). Our own survey of 107 stars has revealed six of the new planets, one of which (around 16 Cygni B) was independently discovered by Bill Cochran and Artie Hatzes (McDonald Observatory, University of Texas). The most recent planet discovery, around Rho Coronae Borealis, was made by a team of nine astronomers led by Robert Noyes of the Harvard-Smithsonian Center for Astrophysics.

The properties of these eight planets are summarized in the accompanying table and diagram. We include one additional object, discovered by David Latham and collaborators, that orbits the star HD 114762 in Coma Berenices. It has a large mass (more than 11 Jupiters), which by some people's definition makes it not exactly a planet. But it's similar to the planet orbiting 70 Virginis, which has a mass of more than 6.5 Jupiters.

We determine the orbital period of each planet directly from the wobble's cycle time. From that period and Kepler's Third Law, the orbit's semimajor axis (the planet's average distance from its star) can quickly be determined. If the velocity varies like a perfect sine wave, we know the orbit is circular. Otherwise, the skewness of the velocity curve can be analyzed to find the eccentricity of the elliptical orbit.

The amplitude of the Doppler variations tell us the planet's minimum mass. If we happen to be seeing the orbit nearly edge on, the star's full orbital velocity is reflected in the Doppler shifts, so the mass we derive will be the planer's true mass. But we don't know our viewing angle, and orbits that are strongly tilted with respect to our line of sight will produce a Doppler shift that is artificially small (due to the diminished amount of approach and recession). So the mass we infer will be smaller than the truth. Thus, we can only determine the mass of the planet multiplied by the sine of the orbit's inclination angle (Msini), where i remains unknown. Luckily this problem is not too great. For randomly oriented orbit planes, the true mass of a planet will be, on average, [Pi]/2 times the minimum mass we determine. So the true planet mass is very probably within twice the value we measure.

The Beasts in the Planetary Zoo

The extrasolar planets discovered around sunlike stars have surprised more than a few astronomers. They all have roughly Jupiter-like masses, with values of Msini between 0.4 and 12 Jupiters. Of course these stars could also have smaller planets that are currently beyond our detection limits. But not one of the 107 sunlike stars on our Lick Observatory observing list has a companion with a minimum mass larger than 6.8 Jupiters. Any companion having a mass of 10 to 80 Jupiters would have been very easily detected. Apparently, nature rarely makes planets much larger than our Jupiter. Why solar systems fail to have super-Jupiters in any great numbers remains unclear.

What's more astonishing is that five of these Jupiter-mass planets orbit remarkably close to their host starts, within 0.25 astronomical unit. This is much closer than Mercury orbits the Sun (0.39 a.u.). The Doppler search technique selects for planets in such tight orbits, because their orbital velocities are very fast. But no one expected to find any giant planet so close to a star, and no one knows for sure how they could have gotten there.

By comparison, the planet orbiting 47 Ursae Majoris seems more similar to our Jupiter. It has a minimum of 2.4 Jupiter masses, its orbit is almost circular, and its orbital radius of 2.1 a.u. corresponds to the inner edge of our asteroid belt. Such a planet, if placed in the solar system, would look like Jupiter's big brother.

Also puzzling are the three planetary companions that have dramatically elliptical orbits. The companions to 16 Cygni B, 70 Virginis, and HD 114762 in Coma Berenices have eccentricities of 0.57, 0.40, and 0.34, respectively. (The largest eccentricities in our solar system, for Mercury and Pluto, are about 0.2; all other planets have nearly circular orbits with eccentricities less than 0.1). Why do some stars have planets in such strongly elliptical orbits?

Indeed, how generic are circular planetary orbits? The arrangement of our own solar system has always seduced theorists into including just enough physical processes in their calculations to yield - ta-da! - circular orbits as "predictions" for solar systems everywhere. Perhaps it's dangerous to have an answer in the back of the book.

When the new discoveries showed that ellipses happen, theorists scurried back to the blackboards and returned quickly with some beautifully sensible elaborations on standard planet-formation theory. Pawel Artymowicz (University of Stockholm) and Pat Cassen (NASA/Ames Research Center) have reanalyzed the birth of planets in the kind of protoplanetary disks that are observed swirling around newborn sunlike stars.

Their calculations show that protoplanets exert gravitational forces on the disk that launch spiral density waves, not unlike tightly wound arms in spiral galaxies. These spiral waves in turn exert subtle gravitational forces back on the planet, pulling it away from pure circular motion. In the course of a million years the departures from circular motion can build up to be quite significant, leading to eccentric orbits.

Further calculations are needed to determine whether eccentricities as high as 0.6 can really result from these spiral waves. Nonetheless, this theory echoes with the wild historical past of planet hunting. The first photographs of "spiral nebulae" (now known to be galaxies), taken by James Keeler at Lick Observatory in the early 1900s, were originally interpreted as planetary systems caught in the act of formation.

A second theory for the large orbital eccentricities seems equally likely. Suppose, for a moment, that our Saturn had grown to a larger mass than it actually did when the solar system was condensing around the newborn Sun. After all, what caused Saturn to stop growing at 1/3 of a Jupiter mass? No one really knows. Indeed, all four giant planets in our solar system might have grown to larger masses if our protoplanetary disk had contained more material or had lasted longer. In that case our solar system would have started out with four true heavyweights, each exerting greater gravitational force on the others than it does now. They would perturb each others' orbits until they crisscrossed. The planets would then gravitationally slingshot around each other. Several researchers have computed this cataclysmic chaos to be the inevitable fate of multiple Jupiters.

These slingshots produce chaotic, unpredictable final orbits almost like the break of balls in billiards. The real planetary pool table, however, is not flat, but has a central depression produced by the gravity of the host star. The scattered Jupiters careen around the depression for a while in wacky interacting orbits. Eventually the situation stabilizes when planets collide and merge, fall into the star, or are flung out of the system entirely. What's left will be just a few planets in eccentric orbiats that are far enough apart not to affect each other much. This model can explain even the most extreme eccentricities seen among the extrasolar planets. If it's what really happened at 16 Cygni B, 70 Virginis, and HD 114762, we should eventually detect other remaining planetary billiard balls in these systems, presumably orbiting much farther out from each star.

A variation on this "gravitational scattering" theory is that a binary star system would perturb planets orbiting tightly around either star, as worked out nicely by Matt Holman (University of Toronto) and collaborators. Indeed, 16 Cygni B is accompanied at a great distance (at least 880 a.u.) by the more massive star 16 Cygni A, the possible "nemesis" responsible for the large eccentricity of the planet around star B.

Mysterious Short-Period Planets

The most mysterious and controversial new planets in the menagerie are the five "hot Jupiters" in circular orbits closer than 0.25 a.u. from their stars. These are the companions of Tau Bootis, 51 Pegasi, Upsilon Andromedae, Rho(1) Cancri, and Rho Coronae Borealis. Their orbital periods are 3.3, 4.2, 4.6, 14.6, and 39.6 days, respectively, compared to Mercury's 88-day orbit around the Sun. The planets' minimum masses range from 0.45 Jupiter (for 51 Peg) to 3.7 Jupiters (for Tau Bootis). All five have essentially circular orbits.

How did these giant planets end up so close to their stars? They starkly contradict standard solar-system formation theory, which dictates that giant planets form at least several a.u. from the host star. There in the cool, outer regions of a protoplanetary disk, small ice grains of frozen water and methane can survive, collide, stick together, and grow to become planetary embryos, constituting the cores of giant planets. If this scenario is correct, how did these five extrasolar giant planets arrive so absurdly close to their stars, where ices can't survive?

There are two possibilities. Either these planets formed right there, in situ, with abundant rocky grains obviating any need for ices, or they formed farther out and migrated inward. The current best theory comes from Douglas Lin, Peter Bodenheimer, and Derek Richardson (University of California at Santa Cruz and University of Washington), borrowing heavily from the brilliant work of William Ward (JPL). In their concept, a protoplanet in its natal disk must migrate inward - for two reasons.

First, the disk material itself is draining ("accreting") into the young star due to viscosity (fluid friction) within the disk. This general flow will drag protoplanets with it. Young solar-type stars do show clear spectral signs of accretion in their excess ultraviolet and infrared emissions. So inward migration can hardly be avoided.

Second, protoplanets will resonate with the spiral waves they set up in the disk, causing them to lose angular momentum and fall inward even if the disk were not itself draining inward. Inevitably, then, protoplanets will be dragged toward the stellar furnace.

This inward planet migration raises enormous conflicts, as yet unresolved. The five short-period planets apparently halted their inward migration and got "parked" in orbits just short of destruction. How did they stop? Second, how did Jupiter and the other giants in our own solar system avoid migrating inward? Why aren't all giant planets found close to their stars?

Douglas Lin actually predicted, before the discovery of the 51-Pegasi type planets, that our own Jupiter should have suffered such a migrational demise. He postulated that previous Jupiters had indeed formed at 5 a.u. and met their fate in the Sun. Our current Jupiter and the rest of the Sun's familiar planets were only the last in line, left stranded when the last of the protoplanetary disk cleared out.

Some preliminary resolutions to this migration catastrophe are emerging. The protoplanetary disks around young sun-like stars probably develop a hole or a gap in their centers with a diameter about 10 times as large as the star itself. This central hole is cleared by the spinning star's magnetic field, which entrains the hot gas (especially the ions and electrons), either flinging it outward or forcing it to flow along magnetic field lines down onto the star. These "doughnut holes" are actually inferred observationally from the lack of near-infrared emission in some young stars.

The hole provides a safe parking spot for the inwardly migrating planet. Once it emerges from the inner edge of the disk, the planet is no longer dragged farther inward. A protoplanet may be fortunate enough to remain parked in the hole during the lifetime of the disk, thus allowing it to orbit there happily ever after.

The magnetic-gap parking mechanism cannot explain the planets around Rho(1) Cancri or Rho Cot Bor, which orbit at 0.11 and 0.23 a.u., respectively. These distances are too great for magnetic fields to clear a hole. So we are left with a mystery. If orbital migration brought Rho(1) Cancri and Rho Cot Bor inward, why didn't they continue migrating? Maybe they were left stranded when the protoplanetary disk dissipated away. Or maybe they formed right where they are, without any orbital migration at all. Perhaps especially massive protoplanetary disks carry enough non-icy material (rock dust and iron dust) to form giant planetary cores.

Why doesn't our solar system have a giant planet in close? Perhaps Jupiter formed near the end of the lifetime of our protoplanetary disk, as the viscosity waned and died out. Or perhaps our disk never had enough gas and dust to have any viscosity in the first place, so it stayed put until the heat of the newborn Sun blew it away. Indeed, protoplanetary disks are known to come in a wide range of masses - from a few Jupiter masses up to hundreds, as deduced from the radio emission by the dust. So the diversity of observed planets may represent a sequence of protoplanetary disk masses or lifetimes, and perhaps chemical compositions.

Planets or Brown Dwarfs?

How can we be sure that the eight companions discovered by the Doppler approach are really "planets" as such, rather than extremely low-mass brown dwarfs? A brown dwarf is defined as a star with too little mass to shine by its own nuclear power, namely less than 0.08 solar mass or about 80 Jupiters.

In today's parlance, the difference between a "brown dwarf" and a "planet" with the same mass in the same orbit depends on how the body formed. A planet grows from dust and gas accreting in a circumstellar disk. A brown dwarf forms the way a star does, by a fragment of a gas cloud collapsing in on itself before the gas has the chance to develop a protoplanetary disk around another star.

Some good evidence bearing on this question comes from the distribution of the newly discovered objects' masses. The graph at upper right shows the occurrence rate of low-mass companions discovered around solar-type stars, brown dwarfs or otherwise. Many of these were found by Michel Mayor and his colleagues in Switzerland and France.

The bin showing the highest rate of companions is the one representing the lowest masses, from 0 to 5 Jupiters. This is absolutely remarkable, because the lowest-mass companions are the most difficult to detect with the Doppler technique (or any other). In fact, we are probably missing many companions having masses below 5 Jupiters because they induce such low Doppler shifts in their stars. But we know that we are hardly missing any companions between 20 and 70 Jupiter masses, because their Doppler amplitude would be so strong.

Clearly there are two populations of low-mass companions to sunlike stars. There is a spotty population having masses between 10 and 70 Jupiters. Some would consider these to be likely brown dwarfs. Then there is the tall peak from 0 to 5 Jupiters. This discontinuous spike suggests a separate class of companions, which may have formed by a different method. The term "planetary" seems quite appropriate for these. One reason is that three of them have circular orbits in situations where the orbits could not have been circularized by tides from the star (47 Ursae Majoris, Rho Cor Bor, and Rho(1) Cancri).

Is Our Solar System a Freak?

With only eight extrasolar planets confirmed and having solid orbital parameters, we lack a good statistical sample against which to compare our own solar system. For example, all the extrasolar planets found so far by the Doppler technique have orbital periods less than 3 years. This is not a reflection of planetary systems in general but rather is due to the limited duration (about 10 years) of the hunts to date. With time and improved Doppler precision, more planets may be found in slower, longer orbits farther from their stars.

However, the meager crop of eight are already sending us a warning message. Our solar system may not be the norm. Suppose that gravitational scattering of planets is common in newborn solar systems. After all, many signs give testimony to heavy bombardment by bodies in crisscrossing orbits when our solar system was young. The cratered face of the Moon and other bodies, the Moon's very formation, and the extreme tilt of the axis of Uranus tell of a violent, impact-racked early solar system. During the first 100 million years it is estimated that there was a new bright comet in the sky of Earth every week. The neat, clean, race-track orbits of today's middle-aged solar system are the crash survivors from the bumper-car era of its reckless youth. There may have been 10 or 11 planets in our solar system during its first 500 million years. We may be unusually lucky that small planets like Earth weren't flung away by giants falling into eccentric orbits, as may have happened with 16 Cygni B.

An outcome of tidy, circular orbits may require special starting conditions. The near-circular orbit of our largest planet, Jupiter, actually promotes the stability of circular orbits among the other eight planets, simulations have found. If Jupiter were in an eccentric orbit, Earth and Mars would likely have been flung out of the solar system long ago. There would be no terrestrial planet in the habitable zone and no Sky & Telescope readers to reflect on this fact. The existence of intelligent life may depend on both Jupiter and Earth being in mutually stable, circular orbits. So of course we will find ourselves in such a system, no matter how unusual it may be. The claim that all planetary orbits must be similar to ours seems like a circular argument.

Planetary Diversity

The handful of planets discovered so far around sunlike stars show orbital properties profoundly more diverse than those found in our solar system. Even the relatively normal-seeming planets of 47 Ursae Majoris and 16 Cygni B have masses more than 1.6 Jupiters, showing that our own Jupiter is not the largest planet that nature can form. The Doppler shifts of 70 Virginis and 16 Cygni B vary in nonsinusoidal cycles that are clearly "Keplerian," the result of bodies orbiting in ellipses according to Kepler's laws. Isaac Newton, were he resurrected, could examine these Doppler data and immediately recognize pure orbital motion stemming from his laws of gravity. Planets are surely orbiting those stars, and they have eccentricities greater than any major planets in our solar system.

It took humanity more than 2,200 years to develop the technology proving that Epicurus was right about the existence of many worlds. The next round of questions should be answered much more quickly. What fraction of stars have planets? What fraction of planetary systems resemble ours? And do small, Earth-sized planets commonly occur in a star's habitable zone, where temperatures are right to allow life?

We continue to monitor 120 stars from Lick Observatory. In July 1996 we began a second Doppler survey of 400 stars using the 10-meter Keck telescope, and in September 1997 we began a survey of 150 stars in the Southern Hemisphere with the 3.9-meter Anglo-Australian Telescope. Michel Mayor and Didier Queloz have recently tripled the size of their Northern Hemisphere Doppler survey to about 400 stars, and in March 1998 they will begin a Southern Hemisphere survey of 500 more. Within the next five years Doppler surveys of several hundred additional stars will begin at the 9-meter Hobby-Eberly Telescope in Texas, directed by William Cochran.

Also within five years, the two Keck telescopes will be linked as an optical interferometer with sufficient precision to detect extrasolar planets by their stars' positional wobbles. And NASA's space-borne interferometer, if funded, should be able to obtain actual images of Earthlike pale blue dots orbiting like clockwork around stars. By 2010 we should have completed the first true census of planets of nearby stars.

Is our heartwarming solar system a freakish twist in the cosmic script, or is it a run-of-the-mill plot device, with other examples sprinkled liberally throughout the solar neighborhood? We don't know, but we will.

RELATED ARTICLE: A Challenge Shot Down

A challenge to the very existence of the five shortest-period extrasolar planets was offered early last year by David Gray (University of Western Ontario). His high-resolution spectra of 51 Pegasi suggested that the shapes of the star's spectral lines are what change with a 4.2-day period, not their overall Doppler shifts (S&T: May 1997, page 24).This would imply that the photosphere of the star is undergoing a complex, nonradial oscillation of a sort never seen in the Sun nor anticipated theoretically in stars. Gray suggested that the Doppler periodicity of 51 Peg can be explained entirely by surface oscillations, eliminating the supposed planet.

This announcement galvanized several groups of astronomers with high-resolution spectrographs into scrutinizing the spectrum of 51 Pegasi in finer detail to check Gray's result, which was based on sporadic observations. A group of nine astronomers, including Tim Brown, Scott Horner, Sylvain Korzennik, and Robert Noyes, first went after the "Gray effect" in Tau Bootis. This star appears to host the most massive and closest-orbiting "hot Jupiter"; its Doppler-shift cycle is the strongest. Indeed, it's so strong that it could have been discovered many years ago had anyone been looking! Tau Bootis showed no variations in the shapes of its spectral lines at all. Thus, oscillations cannot explain the Doppler shifts, and the existence of a planet remains the only viable conclusion. Indeed, even for 51 Pegasi, the planet interpretation neatly explained the single, clear Doppler period of 4.2 days. No oscillatory "overtones" showed up in the Doppler measurements of 51 Pegasi even after data had been collected spanning hundreds of orbital periods.

Then at the end of 1997 came a report by Timothy Brown (High Altitude Observatory) and his colleagues that their high-quality spectra of 51 Pegasi indicate the shapes of the spectral lines are constant. And in January Gray himself withdrew his claim, based on his own further observations and those by others (see page 18).

GEOFFREY MARCY is a professor of physics and astronomy at San Francisco State University and an adjunct professor at the University of California at Berkeley. R. PAUL BUTLER is a staff astronomer at the Anglo-Australian Observatory.
COPYRIGHT 1998 All rights reserved. This copyrighted material is duplicated by arrangement with Gale and may not be redistributed in any form without written permission from Sky & Telescope Media, LLC.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:planets beyond the solar system
Author:Marcy, Geoffrey; Butler, R. Paul
Publication:Sky & Telescope
Article Type:Cover Story
Date:Mar 1, 1998
Previous Article:A closer look at Hubble's Variable Nebula.
Next Article:30 Doradus: birth of a star cluster.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |