Among the most poignant sights in the heavens are white dwarfs. Although they have a mass comparable to our sun’s, they are among the dimmest of all stars and becoming ever dimmer; they do not follow the usual pattern relating stellar mass to brightness. Astronomers think white dwarfs must not be stars so much as the corpses of stars. Each white dwarf was once much like our sun and shone with the same brilliance. But then it began to run out of fuel and entered its stormy death throes, swelling to 100 times its previous size and brightening 10,000-fold, before shedding its outer layers and shriveling to a glowing cinder the size of Earth. For the rest of eternity, it will sit inertly, slowly fading to blackness. As if this story were not gloomy enough, it gets worse. We and our colleagues have found more than a dozen white dwarfs in our galaxy that are orbited by asteroids, comets and perhaps even planets—entire graveyards of worlds. While the stars were still alive, they rose every day in the skies of these worlds. They gently warmed the soil and stirred the wind. Living organisms may have soaked up their rays. But when the stars died, they vaporized or engulfed and incinerated their inner planets, leaving only the bodies that resided in the chilly outposts. Over time the dwarfs shredded and consumed many of the survivors as well. These decimated systems offer a grim look at the fate of our own solar system when the sun dies five billion years from now. Astronomers have always suspected that planets might orbit stars other than our sun. We imagined, though, that we would find systems much like our own solar system, centered on a star much like the sun. Yet when a flood of discoveries began 15 years ago, it was apparent right away that extrasolar planetary systems can differ dramatically from our solar system. The first example was the sunlike star 51 Pegasi, found to have a planet more massive than Jupiter on an orbit smaller than that of Mercury. As instruments became more sensitive, they found ever stranger instances. The sunlike star HD 40307 hosts three planets with masses between four and 10 Earth masses, all on orbits less than half the size of Mercury’s. The sunlike star 55 Cancri A has no fewer than five planets, with masses ranging from 10 and 1,000 Earth masses and orbital radii ranging from one tenth that of Mercury to about that of Jupiter. Planetary systems imagined in science fiction scarcely compare. The white dwarf systems demonstrate that the stars do not even need to be sunlike. Planets and planetary building blocks can orbit bodies that are themselves no larger than planets. The variety of these systems equals that of systems around ordinary stars. Astronomers hardly expected the ubiquity of planetary systems, their hardiness and the apparent universality of the processes by which they form. Solar systems like our own might not be the most common sites for planets, or even life, in the universe. Phoenix from the Ashes It is sometimes forgotten today, but the first confirmed discovery of any extrasolar planets was around a very unsunlike star: the neutron star PSR 1257+12, an even more extreme type of stellar corpse than a white dwarf. It packs a mass greater than the sun’s into the size of a small asteroid, some 20 kilometers across. The event that created this beast, the supernova explosion of a star 20 times the mass of the sun, was more violent than the demise of a sunlike star, and it is hard to imagine planets surviving it. Moreover, the star that exploded probably had a radius larger than 1 AU (astronomical unit, the Earth-sun distance), which is larger than the orbits of the planets we see today. For both reasons, those planets must have risen up out of the ashes of the explosion. Although supernovae typically eject most of their debris into interstellar space, a small amount remains gravitationally bound and falls back to form a swirling disk around the stellar remnant. Disks are the birthing grounds of planets. Astronomers think our solar system took shape when an amorphous interstellar cloud of dust and gas collapsed under its own weight. The conservation of angular momentum, or spin, kept some of the material from simply falling all the way to the newborn sun; instead it settled into a pancake shape. Within this disk, dust and gas coagulated into planets [see “The Genesis of Planets,” by Douglas N. C. Lin; Scientific American, May 2008]. Much the same process could have occurred in the postsupernova fallback disk. Astronomers discovered the system around PSR 1257+12 by detecting periodic deviations in the timing of the radio pulses it gives off; such deviations arise because the orbiting planets pull slightly on the star, periodically shifting its position and thus altering the distance the pulses must travel. Despite intensive searches of other stars’ pulses, observers know of no other comparable system. Another pulsar, PSR B1620–26, has at least one planet, but it orbits so far from the star that astronomers think it did not form in a fallback disk but rather was captured gravitationally from another star. In 2006, however, NASA’s Spitzer Space Telescope discovered unexpected infrared emission from the neutron star 4U 0142+61. The infrared light might arise from the star’s magnetosphere or from a circumstellar disk. This star formed in a supernova explosion about 100,000 years ago, and it typically takes about a million years or so for planets to agglomerate, so if the radiation does signal the presence of a disk, this system may one day resemble that revolving around PSR 1257+12. Many white dwarfs also have disks, albeit of a somewhat different type: disks that indicate the actual presence of orbiting bodies rather than merely the potential to form them. As with 4U 0142+61, the clue is the unexpected emission of infrared light. The first hint dates to 1987, when one of NASA’s ground-based observatories, the Infrared Telescope Facility on the summit of Mauna Kea in Hawaii, found excess infrared light from the white dwarf G29–38. The spectrum of this excess was that of a body with a temperature of 1,200 kelvins, much cooler than the surface of the star, which is 12,000 kelvins. Initially astronomers thought that the dwarf must be orbited by a second, cooler star. But in 1990 they showed that the infrared emission varied in unison with the star’s own brightness, indicating that it was reflected or reprocessed starlight. The most plausible explanation is a circumstellar disk heated by the star. This star has another peculiar property. Its outermost layers contain heavy elements such as calcium and iron, which is odd because the gravitational field near the surface of a white dwarf is so strong that those elements should sink into the interior. In 2003 one of us (Jura) proposed a simple explanation for both the infrared excess and the presence of heavy elements: the white dwarf recently shredded an asteroid that strayed into its intense gravitational field. A cascade of collisions reduced the debris to an orbiting dust disk, which dribbled onto the star. Asteroids for Dessert Observations have since confirmed this scenario. Astronomers using both ground-based telescopes and the Spitzer telescope have identified some 15 white dwarfs with similar infrared excesses and elemental anomalies. For G29–38 and seven other stars, Spitzer has gone further and identified infrared emission from silicates in the disks. These silicates resemble those in dust particles in our solar system and appear quite different from those in dust in interstellar space. Moreover, although the stars’ outer layers contain heavy elements, they do not contain those elements in equal amounts. They are deficient in volatile elements such as carbon and sodium compared with elements that tend to remain in solid form, such as silicon, iron and magnesium. This elemental pattern matches that of the asteroids and rocky planets of the solar system. Both these facts support the contention that the disks are ground-up asteroids. The disks around white dwarfs are much smaller than the disks that give rise to planets around newborn sunlike stars. Judging from their infrared emission, they extend to only about 0.01 AU and have a mass as low as that of an asteroid 30 kilometers in diameter—a fact consistent with their possible origin in the disintegration of such an object. They are not potential sites of the formation of new planets but rather indicators that some planetary material survived the demise of the star. Theoretical calculations suggest that asteroids and Earth-like planets can escape destruction if they orbit farther than 1 AU. When our sun dies, Mars should make it, but Earth may or may not. To study how parts of a planetary system might endure, two years ago Spitzer observed the white dwarf WD 2226–210. This dwarf is so young that the outer layers of the original sunlike star remain visible as the Helix nebula, one of the best-known planetary nebulae [see “The Extraordinary Deaths of Ordinary Stars,” by Bruce Balick and Adam Frank; Scientific American, July 2004]. Consequently, WD 2226–210 provides the missing link between sunlike stars and older white dwarfs such as G29–38. Around it is a dusty disk at a distance of 100 AU, comparable to the scale of our solar system. That is much farther than disks around other white dwarfs extend—too far, in fact, to consist of asteroids torn up by the dwarf’s gravity. This disk must instead consist of dust released as asteroids and comets collide. Similar debris disks exist around the sun and sunlike stars [see “The Hidden Members of Planetary Systems,” by David R. Ardila; Scientific American, April 2004]. This discovery confirms that when a sunlike star dies, distant asteroids and comets can survive. And if asteroids and comets can survive, planets (which are, if anything, more durable) should be able to survive as well. As WD 2262–210 cools, it will give off less light to illuminate the dust, and the distant belt of asteroids and comets will fade into invisibility. But occasionally one of its members may wander close enough to the white dwarf to be shredded. Starlets A third type of nonsunlike star that might host planets is the brown dwarf. Brown dwarfs are very different from white dwarfs, despite the similar names. They are not stellar corpses but stellar runts. They form in the same way stars do, but their growth is stunted, leaving them with less than about 8 percent the mass of the sun—the threshold required for a stellar core to become hot and dense enough to ignite sustained nuclear fusion. The most they manage is a feeble infrared glow as they radiate away the heat they accumulated during their formation (and perhaps a brief early period of fusion). Over the past 15 years astronomical surveys have found hundreds of brown dwarfs, and the least massive of them is scarcely heavier than a giant planet. Astronomers have found that these bodies, even the smallest among them, can have disks and therefore perhaps planets as well [see “The Mystery of Brown Dwarf Origins,” by Subhanjoy Mohanty and Ray Jayawardhana; Scientific American, January 2006]. The possibility of planets is supported by observations showing that brown dwarf disks undergo a series of systematic changes—including a drop in the prominence of the infrared emission from silicates—attributable to coagulation of the dust particles. The same changes also occur in disks around larger stars and signal the growth of planetary building blocks. The brown dwarf disks are too meager for planets as large as Jupiter to form but contain plenty of material for a Uranus or Neptune. Some astronomers have claimed the discovery of planets that formed around brown dwarfs, but none of these claims is definitive. In short, astronomers have found planets around at least one neutron star; asteroids and comets around more than a dozen white dwarfs; and evidence for the early stages of planet formation around brown dwarfs. Ultimately, the study of these and other extrasolar systems has two goals: First, astronomers hope to learn more about our own solar system, particularly about its evolution and large-scale structure, features that are hard to discern from our limited temporal and spatial perspective. We also hope to place our solar system in its context. Is it average or an outlier? Despite the diversity of planetary systems, do they follow some common pathways in their formation? The similarity between the composition of asteroids in our solar system and of the material that has fallen onto white dwarfs suggests that the answer is yes. The second goal is to determine how widespread life might be in the universe. In our galactic neighborhood, brown dwarfs are roughly as numerous as stars. Might the nearest “star” to our sun be a yet to be discovered brown dwarf? Might the nearest planets to our solar system orbit a brown dwarf? The Wide-field Infrared Survey Explorer (WISE) satellite, which NASA plans to launch at the end of the year, may well discover several brown dwarfs closer than the nearest known star. The formation of terrestrial planets around brown dwarfs would not only extend the range of potential habitats but also lead to the intriguing possibility that the nearest extraterrestrial life may wake up in the morning to a brown dwarf. Similarly, the presence of asteroids and comets around white dwarfs raises the possibility not only that planets can survive the demise of a sunlike star but also that life, if it could adapt to the changing conditions, may hold out in the environs of these dead stars. Perhaps, then, white dwarfs are not such a gloomy sight after all. Glowing in the Dark Astronomers generally detect planets indirectly, by virtue of their effects on the velocity, position or brightness of their host stars. For most of the cases discussed in the article, astronomers focus on one type of indirect sign: the presence of a disk of dust orbiting the star. A so-called protoplanetary disk occurs around newly born stars and is thought to be the site of planet formation. A so-called debris disk occurs around mature stars and is thought to arise from collisions or evaporation of comets and asteroids, thus signaling the likely presence of planets now or in the past. Observers identify both types of disk from how they absorb starlight and reradiate the absorbed energy at infrared wavelengths. NASA’s Spitzer Space Telescope, launched in 2003, has proved to be a veritable disk discovery machine. Its large field-of-view infrared cameras can capture hundreds of stars in a single image and pinpoint those with evidence of disks for further study. Spitzer builds on the successes of past infrared telescopes, such as the Infrared Astronomical Satellite (IRAS) mission in the 1980s and the European Space Agency’s Infrared Space Observatory (ISO) in the mid-1990s. Unlike IRAS, which was an all-sky survey, Spitzer points at specific celestial bodies for intensive study, and the five-year-plus lifetime of its liquid-helium coolant far exceeds that of any previous mission. The telescope has studied everything from extrasolar planets to galaxies in the early universe. The coolant is now running out, and the telescope will soon start to warm from nearly absolute zero to 30 kelvins. Even so, it will be able to operate at the short-wavelength end of the infrared band through at least the middle of 2011. Taking up the slack will be the newly launched Herschel Space Observatory and the James Webb Space Telescope (JWST), planned for launch in 2013. Note: This article was originally printed with the title, “Improbable Planets”.
As if this story were not gloomy enough, it gets worse. We and our colleagues have found more than a dozen white dwarfs in our galaxy that are orbited by asteroids, comets and perhaps even planets—entire graveyards of worlds. While the stars were still alive, they rose every day in the skies of these worlds. They gently warmed the soil and stirred the wind. Living organisms may have soaked up their rays. But when the stars died, they vaporized or engulfed and incinerated their inner planets, leaving only the bodies that resided in the chilly outposts. Over time the dwarfs shredded and consumed many of the survivors as well. These decimated systems offer a grim look at the fate of our own solar system when the sun dies five billion years from now.
Astronomers have always suspected that planets might orbit stars other than our sun. We imagined, though, that we would find systems much like our own solar system, centered on a star much like the sun. Yet when a flood of discoveries began 15 years ago, it was apparent right away that extrasolar planetary systems can differ dramatically from our solar system. The first example was the sunlike star 51 Pegasi, found to have a planet more massive than Jupiter on an orbit smaller than that of Mercury. As instruments became more sensitive, they found ever stranger instances. The sunlike star HD 40307 hosts three planets with masses between four and 10 Earth masses, all on orbits less than half the size of Mercury’s. The sunlike star 55 Cancri A has no fewer than five planets, with masses ranging from 10 and 1,000 Earth masses and orbital radii ranging from one tenth that of Mercury to about that of Jupiter. Planetary systems imagined in science fiction scarcely compare.
The white dwarf systems demonstrate that the stars do not even need to be sunlike. Planets and planetary building blocks can orbit bodies that are themselves no larger than planets. The variety of these systems equals that of systems around ordinary stars. Astronomers hardly expected the ubiquity of planetary systems, their hardiness and the apparent universality of the processes by which they form. Solar systems like our own might not be the most common sites for planets, or even life, in the universe.
Phoenix from the Ashes It is sometimes forgotten today, but the first confirmed discovery of any extrasolar planets was around a very unsunlike star: the neutron star PSR 1257+12, an even more extreme type of stellar corpse than a white dwarf. It packs a mass greater than the sun’s into the size of a small asteroid, some 20 kilometers across. The event that created this beast, the supernova explosion of a star 20 times the mass of the sun, was more violent than the demise of a sunlike star, and it is hard to imagine planets surviving it. Moreover, the star that exploded probably had a radius larger than 1 AU (astronomical unit, the Earth-sun distance), which is larger than the orbits of the planets we see today. For both reasons, those planets must have risen up out of the ashes of the explosion.
Although supernovae typically eject most of their debris into interstellar space, a small amount remains gravitationally bound and falls back to form a swirling disk around the stellar remnant. Disks are the birthing grounds of planets. Astronomers think our solar system took shape when an amorphous interstellar cloud of dust and gas collapsed under its own weight. The conservation of angular momentum, or spin, kept some of the material from simply falling all the way to the newborn sun; instead it settled into a pancake shape. Within this disk, dust and gas coagulated into planets [see “The Genesis of Planets,” by Douglas N. C. Lin; Scientific American, May 2008]. Much the same process could have occurred in the postsupernova fallback disk.
Astronomers discovered the system around PSR 1257+12 by detecting periodic deviations in the timing of the radio pulses it gives off; such deviations arise because the orbiting planets pull slightly on the star, periodically shifting its position and thus altering the distance the pulses must travel. Despite intensive searches of other stars’ pulses, observers know of no other comparable system. Another pulsar, PSR B1620–26, has at least one planet, but it orbits so far from the star that astronomers think it did not form in a fallback disk but rather was captured gravitationally from another star.
In 2006, however, NASA’s Spitzer Space Telescope discovered unexpected infrared emission from the neutron star 4U 0142+61. The infrared light might arise from the star’s magnetosphere or from a circumstellar disk. This star formed in a supernova explosion about 100,000 years ago, and it typically takes about a million years or so for planets to agglomerate, so if the radiation does signal the presence of a disk, this system may one day resemble that revolving around PSR 1257+12.
Many white dwarfs also have disks, albeit of a somewhat different type: disks that indicate the actual presence of orbiting bodies rather than merely the potential to form them. As with 4U 0142+61, the clue is the unexpected emission of infrared light. The first hint dates to 1987, when one of NASA’s ground-based observatories, the Infrared Telescope Facility on the summit of Mauna Kea in Hawaii, found excess infrared light from the white dwarf G29–38. The spectrum of this excess was that of a body with a temperature of 1,200 kelvins, much cooler than the surface of the star, which is 12,000 kelvins.
Initially astronomers thought that the dwarf must be orbited by a second, cooler star. But in 1990 they showed that the infrared emission varied in unison with the star’s own brightness, indicating that it was reflected or reprocessed starlight. The most plausible explanation is a circumstellar disk heated by the star.
This star has another peculiar property. Its outermost layers contain heavy elements such as calcium and iron, which is odd because the gravitational field near the surface of a white dwarf is so strong that those elements should sink into the interior. In 2003 one of us (Jura) proposed a simple explanation for both the infrared excess and the presence of heavy elements: the white dwarf recently shredded an asteroid that strayed into its intense gravitational field. A cascade of collisions reduced the debris to an orbiting dust disk, which dribbled onto the star.
Asteroids for Dessert Observations have since confirmed this scenario. Astronomers using both ground-based telescopes and the Spitzer telescope have identified some 15 white dwarfs with similar infrared excesses and elemental anomalies. For G29–38 and seven other stars, Spitzer has gone further and identified infrared emission from silicates in the disks. These silicates resemble those in dust particles in our solar system and appear quite different from those in dust in interstellar space. Moreover, although the stars’ outer layers contain heavy elements, they do not contain those elements in equal amounts. They are deficient in volatile elements such as carbon and sodium compared with elements that tend to remain in solid form, such as silicon, iron and magnesium. This elemental pattern matches that of the asteroids and rocky planets of the solar system. Both these facts support the contention that the disks are ground-up asteroids.
The disks around white dwarfs are much smaller than the disks that give rise to planets around newborn sunlike stars. Judging from their infrared emission, they extend to only about 0.01 AU and have a mass as low as that of an asteroid 30 kilometers in diameter—a fact consistent with their possible origin in the disintegration of such an object. They are not potential sites of the formation of new planets but rather indicators that some planetary material survived the demise of the star. Theoretical calculations suggest that asteroids and Earth-like planets can escape destruction if they orbit farther than 1 AU. When our sun dies, Mars should make it, but Earth may or may not.
To study how parts of a planetary system might endure, two years ago Spitzer observed the white dwarf WD 2226–210. This dwarf is so young that the outer layers of the original sunlike star remain visible as the Helix nebula, one of the best-known planetary nebulae [see “The Extraordinary Deaths of Ordinary Stars,” by Bruce Balick and Adam Frank; Scientific American, July 2004].
Consequently, WD 2226–210 provides the missing link between sunlike stars and older white dwarfs such as G29–38. Around it is a dusty disk at a distance of 100 AU, comparable to the scale of our solar system. That is much farther than disks around other white dwarfs extend—too far, in fact, to consist of asteroids torn up by the dwarf’s gravity. This disk must instead consist of dust released as asteroids and comets collide. Similar debris disks exist around the sun and sunlike stars [see “The Hidden Members of Planetary Systems,” by David R. Ardila; Scientific American, April 2004].
This discovery confirms that when a sunlike star dies, distant asteroids and comets can survive. And if asteroids and comets can survive, planets (which are, if anything, more durable) should be able to survive as well. As WD 2262–210 cools, it will give off less light to illuminate the dust, and the distant belt of asteroids and comets will fade into invisibility. But occasionally one of its members may wander close enough to the white dwarf to be shredded.
Starlets A third type of nonsunlike star that might host planets is the brown dwarf. Brown dwarfs are very different from white dwarfs, despite the similar names. They are not stellar corpses but stellar runts. They form in the same way stars do, but their growth is stunted, leaving them with less than about 8 percent the mass of the sun—the threshold required for a stellar core to become hot and dense enough to ignite sustained nuclear fusion. The most they manage is a feeble infrared glow as they radiate away the heat they accumulated during their formation (and perhaps a brief early period of fusion). Over the past 15 years astronomical surveys have found hundreds of brown dwarfs, and the least massive of them is scarcely heavier than a giant planet.
Astronomers have found that these bodies, even the smallest among them, can have disks and therefore perhaps planets as well [see “The Mystery of Brown Dwarf Origins,” by Subhanjoy Mohanty and Ray Jayawardhana; Scientific American, January 2006]. The possibility of planets is supported by observations showing that brown dwarf disks undergo a series of systematic changes—including a drop in the prominence of the infrared emission from silicates—attributable to coagulation of the dust particles. The same changes also occur in disks around larger stars and signal the growth of planetary building blocks. The brown dwarf disks are too meager for planets as large as Jupiter to form but contain plenty of material for a Uranus or Neptune. Some astronomers have claimed the discovery of planets that formed around brown dwarfs, but none of these claims is definitive.
In short, astronomers have found planets around at least one neutron star; asteroids and comets around more than a dozen white dwarfs; and evidence for the early stages of planet formation around brown dwarfs. Ultimately, the study of these and other extrasolar systems has two goals: First, astronomers hope to learn more about our own solar system, particularly about its evolution and large-scale structure, features that are hard to discern from our limited temporal and spatial perspective. We also hope to place our solar system in its context. Is it average or an outlier? Despite the diversity of planetary systems, do they follow some common pathways in their formation? The similarity between the composition of asteroids in our solar system and of the material that has fallen onto white dwarfs suggests that the answer is yes.
The second goal is to determine how widespread life might be in the universe. In our galactic neighborhood, brown dwarfs are roughly as numerous as stars. Might the nearest “star” to our sun be a yet to be discovered brown dwarf? Might the nearest planets to our solar system orbit a brown dwarf? The Wide-field Infrared Survey Explorer (WISE) satellite, which NASA plans to launch at the end of the year, may well discover several brown dwarfs closer than the nearest known star. The formation of terrestrial planets around brown dwarfs would not only extend the range of potential habitats but also lead to the intriguing possibility that the nearest extraterrestrial life may wake up in the morning to a brown dwarf.
Similarly, the presence of asteroids and comets around white dwarfs raises the possibility not only that planets can survive the demise of a sunlike star but also that life, if it could adapt to the changing conditions, may hold out in the environs of these dead stars. Perhaps, then, white dwarfs are not such a gloomy sight after all.
Glowing in the Dark Astronomers generally detect planets indirectly, by virtue of their effects on the velocity, position or brightness of their host stars. For most of the cases discussed in the article, astronomers focus on one type of indirect sign: the presence of a disk of dust orbiting the star. A so-called protoplanetary disk occurs around newly born stars and is thought to be the site of planet formation. A so-called debris disk occurs around mature stars and is thought to arise from collisions or evaporation of comets and asteroids, thus signaling the likely presence of planets now or in the past.
Observers identify both types of disk from how they absorb starlight and reradiate the absorbed energy at infrared wavelengths. NASA’s Spitzer Space Telescope, launched in 2003, has proved to be a veritable disk discovery machine. Its large field-of-view infrared cameras can capture hundreds of stars in a single image and pinpoint those with evidence of disks for further study.
Spitzer builds on the successes of past infrared telescopes, such as the Infrared Astronomical Satellite (IRAS) mission in the 1980s and the European Space Agency’s Infrared Space Observatory (ISO) in the mid-1990s. Unlike IRAS, which was an all-sky survey, Spitzer points at specific celestial bodies for intensive study, and the five-year-plus lifetime of its liquid-helium coolant far exceeds that of any previous mission. The telescope has studied everything from extrasolar planets to galaxies in the early universe.
The coolant is now running out, and the telescope will soon start to warm from nearly absolute zero to 30 kelvins. Even so, it will be able to operate at the short-wavelength end of the infrared band through at least the middle of 2011. Taking up the slack will be the newly launched Herschel Space Observatory and the James Webb Space Telescope (JWST), planned for launch in 2013.
Note: This article was originally printed with the title, “Improbable Planets”.
