Chapter 2
Planets from Dust
Unraveling the Birth of Solar Systems
Since there is no way to go back in time and see how our own solar system formed, astronomers have to find clues to its origin by other means. Detailed observations of stellar nurseries reveal the characteristics of young suns and their surroundings, including disks of gas and dust that presumably turn into planetary systems. Sophisticated computer simulations, based on our understanding of physical laws, can follow the collapse, under the influence of gravity, of a gas cloud into a star. Experiments with dust balls show how tiny grains stick together to make bigger clumps in disks girdling young stars. Other evidence comes from the solar system it-self—the ordering of planetary orbits and the makeup of comets and meteorites. By pulling together all these clues, astronomers are now able to decipher many of the critical stages in the birth of stars and their planetary retinues.
Cosmic Cradles
The constellation Orion the Hunter is easy to spot in the winter evening sky. Slightly below the three stars that make up the Hunter’s belt, just about in the middle of his sword, is a giant cloud of gas and dust known as the Orion Nebula. Illuminated by several hot stars that make up the famous Trapezium at its core, it is quite a pretty sight through a pair of binoculars or a small telescope.
Under good sky conditions, the nebula is visible to the naked eye as a fuzzy patch. Thus it is somewhat surprising that no records of it exist until the advent of the telescope. Galileo surveyed Orion’s belt region in 1609 and catalogued several stars, but he did not report seeing the nebula. Its discovery is credited to a French lawyer by the name of Nicholas-Claude Fabri de Peiresc, who observed it with his telescope a year later. The following year, Jesuit astronomer Johann Baptist Cysatus found it independently and compared the nebula to a comet he had observed around the same time. Giovanni Batista Hodierna of Italy included the first known drawing of the Orion Nebula in a catalog published in 1654.
In the two centuries that followed, a string of famous observers—including Christian Huygens, Charles Messier, and William Herschel—carefully studied it. Messier included it as the forty-second object in his catalog of nebulae so that it would not be confused as a comet. In 1774, Herschel, the discoverer of planet Uranus, described it as “an unformed fiery mist, the chaotic material of future suns”—words that sound prophetic in hindsight—but his claim that the nebula’s appearance changed over a period of several years is almost certainly spurious. Between 1858 and 1863, George Bond of Harvard College Observatory produced one of the most beautiful drawings of the nebula ever. Henry Draper, a professor of physiology at New York University, took the first photograph of it using an 11-inch telescope in 1880. Today’s professional astronomers turn their powerful telescopes, on the ground and in space, toward the Orion Nebula, too, because it holds clues to the birth of stars and even of solar systems like our own.
The Orion Nebula, at a distance of roughly 1,500 light-years, is one of the stellar nurseries nearest to us. At its heart is a dense cluster of about three thousand very young stars, barely a million years old, packed into a volume just a few light-years across. Only a handful of the brightest stars in this Trapezium cluster are visible in optical images. The rest are still obscured by dust grains in their natal cloud. However, at longer infrared and millimeter wavelengths, we can peer through the dust and see the true majesty of a thousand suns recently born. (It’s for the same reason that you can hear your favorite radio station on a foggy day even if you can’t see more than a few meters in front of you.) In fact, the nebula is part of a much bigger complex of “molecular clouds” that covers much of the Orion constellation and also includes other celestial objects familiar to stargazers, such as the Horsehead Nebula and Barnard’s Loop. We can see only those parts that shine by the light of newborn stars. The rest of the cloud is extremely cold and dark and is made up mostly of hydrogen molecules (hence the name “molecular cloud”) and a smidgen of dust.
The study of stellar nurseries began some two hundred years ago with the work of William Herschel. Born in Hanover, Germany as one of ten children, he traveled to England at age eighteen with the local regimental band, in which he played the oboe. He moved to England the following year, where his sister Caroline joined him later. A versatile musician, Herschel also played the cello and the organ and composed two dozen symphonies and numerous concertos. It was while working in Bath as a church organist that he became interested in astronomy. After reading books on optics, he started grinding and polishing his own mirrors for telescopes he built in his backyard. With each new instrument, he eagerly scanned the heavens. Soon, this self-taught astronomer became the world’s leading observer.
During his systematic surveys of the sky, done with the help of sister Caroline, Herschel noticed that some small patches of the sky were remarkably devoid of stars. He called them “holes in the heavens.” It took almost a century and the development of photography before astronomers understood what these “holes” were. When the American astronomer Edward Emerson Barnard took long-exposure photographs of these dark areas in the 1920s, they appeared more like clouds than holes. Around the same time, other astronomers found evidence for dust in space between the stars. That convinced Barnard that Herschel’s “holes” were in fact dusty clouds that blocked the light from stars behind them.
Over a decade starting in 1948, the 1.2-meter Schmidt telescope on Mount Palomar in Southern California carried out a photographic survey of the entire northern sky. In 1962, after carefully examining all the Palomar plates, Beverly Lynds, then at the University of Arizona, published a catalog of 1,801 dark clouds like the ones Herschel had first noticed. More recently, astronomers have compiled a list of dark clouds in the southern hemisphere as well, bringing the total to almost 3,000.
These clouds are among the coldest objects in the Galaxy. Their temperature is about –260 degrees Celsius. The clouds are about 99 percent gas and 1 percent dust grains—yes, the same sort of dust that settles in your room, but these particles are even smaller. Most of the gas is hydrogen, but astronomers have also detected carbon monoxide, water, and ammonia in interstellar clouds, using radio telescopes that can detect emission from those molecules.
The biggest clouds, like the one in Orion, are called giant molecular clouds. One cloud can extend for hundreds of light-years and have a mass as much as 10 million times that of the Sun. Because the mass is spread out over such a large volume, these clouds are in fact very tenuous; typically there is only one molecule per cubic centimeter of volume in a large cloud. There are also medium-size clouds, some of the nearest of which are found in the constellation Taurus, with masses that are a few thousand times the mass of the sun. The smallest clouds, called Bok globules after astronomer Bart Bok who first pointed them out in 1947, are just a few solar masses and extend barely one light-year across.
It is inside these clouds of gas and dust that new stars are born. The process begins with particles in some small region of the cloud coming together to form a little clump or a “core.” Random motions of gas molecules may bring enough mass together to make the initial clump. Or an external disturbance, such as a passing shock wave from a nearby supernova, may induce a portion of the cloud to contract. In either case, once a seed core forms, its gravity pulls in more and more material, and it grows more massive. The core eventually becomes so massive that it collapses—or shrinks—under its own gravity. As it collapses, the core probably breaks into smaller, spinning blobs. Each blob becomes a dense ball of gas. The shrinking blobs rotate faster and faster, just as a figure skater spins faster and faster as she brings in her arms to conserve angular momentum. As each blob gets denser, it also gets hotter and starts to glow. These glowing blobs are what we call baby stars or “proto-stars.” New stars often come in pairs or triples, the result of the original core fragmenting into two or three clumps. All this happens in a few hundred thousand years—a pretty short time, given that the Sun’s total lifetime is about 10 billion years. As early as the 1920s, the English astronomer James Jeans suggested that a cloud would collapse under its own gravity if it gets too massive. Basically the inward pull of gravity wins over the outward pressure of gas.
Hazy Beginnings
Frank Shu went to first grade three times. Not that he needed remedial help as a child. In fact, at the tender age of four, he was tested into the third grade. But his mother decided it would be best for him to enter first grade with his peers. Within a year, the family moved twice, from mainland China, where he was born in 1943, to Taiwan, and then to the United States. And young Shu attended first grade in each of those three places. “So, I know that material extremely well!” he once told me, chuckling. Perhaps it wasn’t such a bad way to start a life in the academy. Today, Shu holds the prestigious title of University Professor in the University of California system and is widely regarded as one of the world’s foremost experts on star formation. His theoretical work over the past three decades is at the heart of our current understanding of the birth process of stars. In 2009, he received the million-dollar Shaw Prize for his wide-ranging contributions. When he is not doing astronomy, Shu enjoys playing poker and bridge as well as tennis and billiards. He brings a competitive spirit to these pursuits, as I witnessed for myself during a poker game at a conference in Florida a few years ago. One of Shu’s early heroes was Leonardo da Vinci. “Soon after we came to the U.S., my father took us to an exhibition on da Vinci,” he said. “I remember being really impressed by this guy who could draw, who could do science, who invented so many things . . . the definition of a Renaissance man.” Perhaps that early impression has stayed with Shu to this day. He sees much in common between science and art. “The scientist at his or her purest is very similar to the artist,” he explained. “They have common goals; they’re both searching for the truth. The difference is that scientific truth is external truth whereas the truth that a writer or a painter sees is inner truth.”
In 1977, Shu published a seminal paper on star formation, building on previous work by Yale University astronomer Richard Larson and others. In it, he proposed a simple, yet elegant, model showing that cloud cores collapse “inside out,” first forming a small central star onto which rest of the material falls. Because the cloud is spinning, it actually fattens into a disk as it shrinks in size, sort of like how pizza dough makes a pie as it is spun in the air. Therefore, the rest of the cloud material actually falls on to this disk, rather than directly onto the newborn star. Later, material in the disk spirals in toward the baby star. Some of that stuff is shot out from the poles of the baby star as the inner part of the disk rubs against the star’s magnetic field.
Protostars do not yet have sufficient heat and pressure in their cores to ignite nuclear reactions, the energy source of stars. Instead, their glow comes from converting gravitational energy into heat, as they continue to contract.
Until recently, the progression from cloud to proto-star was hidden from astronomers’ view. Visible light does not escape through the dark shroud of dust surrounding the stellar embryos, but radio waves and infrared radiation do. Over the past two decades, with the development of sensitive detectors at these longer wavelengths, astronomers have been able to peer deep into the heart of stellar nurseries and see the early stages of star birth. Researchers such as Phil Myers at the Harvard-Smithsonian Center for Astrophysics have identified hundreds of dense cores in nearby dark clouds by observing the emission of molecules such as ammonia with radio telescopes. Since ammonia molecules (unlike hydrogen molecules) are found in denser parts of the gas, their emission can be used to trace the location, size, and mass of cores. Myers and his colleagues have identified hundreds of cores in nearby dark clouds. Some of the cores appear to be shrinking, or collapsing inward, and harbor strong infrared sources, a sure sign of a newborn star.
Another telltale signature of a protostar is a pair of jets coming out in opposite directions. Twisting magnetic-field lines between the protostar and its disk are believed to be responsible for this spectacular phenomenon: a fraction of the material that spirals in through the disk is shot out from the star’s poles, in opposite directions, perpendicular to the disk plane. Many of the jets contain clumps, suggesting that material is being shot out in machine-gun–like fashion every few decades. Since the jet origin is intimately linked to accretion from the disk, the presence of clumps implies that the disk also dumps material onto the star in spurts. When jets, moving at hundreds of thousands of kilometers per hour, slam into the surrounding interstellar gas, the collision creates a bow shock, similar to that made by a speedboat skimming across a lake. The violent collision heats up the stationary gas, and the result is glowing shock regions known as Herbig-Haro objects, after George Herbig and Guillermo Haro who discovered them in the 1950s.
Figure 2.1. Stages of star birth, from a cold cloud of gas and dust to a mature star accompanied by a full-fledged retinue of planets.
In addition to the jets, which are narrow, hot, and fast moving, many protostars also harbor broader and slower outflows of cold gas. The jets are usually nested inside these vast outflows. While the ionized gas in jets can be seen in optical and near-infrared images, the colder outflows are best detected through the emission of specific molecules at radio wavelengths.
Once the dust settles somewhat (literally and figuratively), the embryonic star becomes visible to optical telescopes, typically at an age of about a million years. During this phase, which astronomers call T Tauri after the prototype in the Taurus star-forming region, the star is prone to violence. There are frequent outbursts of energy, sometimes seen as X-ray fares, and powerful winds. Gigantic starspots, much bigger than sunspots, dot its surface. T Tauri stars are still surrounded by disks, though much less massive than those around protostars. Much of the original disk material has already been accreted into the star or blown away by winds, so the rate of accretion through the disk has slowed to a crawl. As a result, their jets are weaker too.
Eventually, the center of the contracting star reaches high-enough temperatures and pressures to fuse hydrogen nuclei into helium nuclei. Once the nuclear reactions ignite, at a temperature of 10 million Kelvin, the outward pressure of heated gas halts further gravitational contraction. The star achieves a fine balance, or equilibrium, and spends most of its life in this hydrogen-fusing “main sequence” phase. How long a star will stay on the main sequence depends primarily on how massive it is: prodigal high-mass stars burn their fuel much faster than their parsimonious low-mass cousins. The Sun would last 10 billion years on the main sequence, whereas a star with three times its mass would run out of hydrogen twenty times sooner.
Planet Building
When the Hubble Space Telescope turned toward the Orion Nebula in 1992, its images showed that many of the baby stars in Orion are surrounded by dusty disks, seen in silhouette against the bright background of the nebula. It is out of these “leftover” disks that planets form. Therefore, in recent years, astronomers have made significant efforts to understand the frequency and characteristics of these disks. Even before the Hubble images of Orion, researchers had taken a census of protoplanetary disks by less direct means. Since dust particles absorb a star’s light and re-emit it in the infrared, stars with disks would shine brighter at those wavelengths than otherwise. Surveys of nearby star-forming regions had revealed that 50 to 90 percent of very young stars harbor an “infrared excess” consistent with the presence of disks. Still, it was reassuring to see the disks of Orion directly with the Hubble.
Figure 2.2. Dust grains in the disk absorb starlight and emit that heat in the infrared, producing an excess that reveals their presence.
These disks are typically a few hundred astronomical units across, or several times the size of Pluto’s orbit around the Sun. Observations with millimeter-wave telescopes allow astronomers like Anneila Sargent at the California Institute of Technology to estimate how much material they contain: the disk masses range from about 1 percent to 10 percent of the Sun’s mass. That is more than enough to make a planetary system like ours. Even though dust makes up only about 1 percent of the mass—the rest being gas, mostly hydrogen and helium—it emits almost all the infrared radiation. To map gas in the disk, astronomers observe emission from specific molecules, like carbon monoxide, with radio telescopes.
Some of the disks, in places like the Orion Nebula, won’t survive long enough to grow planets. Intense ultraviolet radiation from hot stars in the Trapezium is boiling off material from the surfaces of many “proplyds.” Hubble images show that proplyds are surrounded by a comet-like envelope of evaporating material with its tail pointing away from the nearby hot stars. These disks are losing material at a furious rate of about 0.5 Earth masses per year and will be almost completely evaporated within a few hundred thousand years, probably before planets have a chance to form. Living so close to Trapezium’s bigwigs presents “an environmental hazard to planet formation,” as John Bally of the University of Colorado in Boulder puts it. But elsewhere, in quieter environments like the Taurus clouds, the protoplanetary disks last much longer.
How and when do these disks make planets? That question has been the focus of recent research on many fronts. There is growing evidence that disks evolve into planetary systems within about 10 million years. The saga, as it is generally told, begins with the dust grains in the disk sticking together to make things as big as pebbles. Those pebbles then collide and stick together to build up boulder-size objects called planetesimals, which are in turn the building blocks of planets. Scientists such as George Wetherill at the Carnegie Institution of Washington and Stuart Weidenschilling at the Planetary Science Institute in Tucson have tried to follow this process using computer simulations. These simulations show that making pebbles into boulders is not easy: their gravity is too weak to attract other pebbles, and even if pebbles happen to hit each other, they often shatter into pieces instead of sticking together.
To solve the mystery of how planetesimals grow, Jürgen Blum at the University of Jena in Germany and his colleagues play with dust in their labs. Many of their experiments require almost weightless—or “micro-gravity”—conditions so that effects of the Earth’s gravity do not overshadow weaker molecular processes that they try to investigate. That’s where parabolic fights, in a modified Airbus 300 jet rented by the European Space Agency, come in handy. Flying some 10,000 meters above sea level off the coast of France, the plane climbs upward at a forty-five-degree angle, plunges down in free fall, and then recovers. These roller-coaster fights provide scientists with a series of twenty-second windows of microgravity in which they conduct experiments on how dust particles stick together in specially designed vacuum chambers cooled to –200 degrees Celsius. An ultra-fast camera attached to the chambers takes pictures every few nanoseconds.
Other experiments are conducted in space. In one, performed aboard the space shuttle Discovery in late 1998, astronauts injected micron-size dust grains into a chamber filled with low-pressure gas, to simulate conditions in the protoplanetary disk. Then they took photographs with microscopes to see how the particles interacted. Within minutes, dust grains stuck together to make stringy structures, rather than spherical clumps as some simulations had predicted. The same process would take about a year in the actual disk, where the densities are about a million times lower than in the shuttle experiment. A later experiment, launched aboard an unmanned sounding rocket from the north of Sweden, confirmed these findings and showed that the dust structures grew at an exponentially increasing rate. That’s good news for the first stages of planet growth.
Some experiments show that when these long, thin chains collide at low speeds, they build up fluffy aggregates about the size of large pebbles, but a lot less dense. As the dust balls grow, they tend to sediment, or settle, to the bottom of the jar, thus increasing the chance of collisions and further growth. Blum and his collaborators are also investigating how these dust balls can build up kilometer-size objects. One possibility is that aerodynamic capture (remember, the dust particles are suspended in gas) of small particles by bigger grains could allow the latter to keep growing. To test that idea, Blum’s group bombards a porous “dust cake”—simulating the surface of a small planetesimal—with little dust balls.
Other scientists, like Anders Johansen at Lund University in Sweden, suggest that turbulent motions of gas in protoplanetary disks may help. Their computer simulations show that wherever the gas density is slightly higher, solid particles also tend to gather. Eventually, these over-dense regions contract under gravity to form asteroid-size objects directly.
Once you build up planetesimals, the rest is easier. Larger planetesimals have enough gravity to hold on to almost anything that hit them, so they grow bigger. Smaller planetesimals get stuck onto big ones or they are shattered into dust by collisions. As planetesimals get up to about the size of the Moon, collisions become rarer; but when they do occur they tend to be more violent.
Within a few million years, rocky planets like Earth and Mars would have pretty much reached their final mass.
But, in the standard story, giant planets like Jupiter and Saturn still have ways to grow. Far from the sun, beyond what’s called the snow line, it’s cold enough for ice to form in the disk. The presence of ice means there is a lot more solid material out there for planet making. Starting with solid cores of dust and ice, with perhaps ten times the Earth’s mass, planets build up thick atmospheres by sweeping up vast amounts of gas. That process could take several million years. The question is whether those planets could gather enough gas before the young sun blows it all away. Or could it be that giant planets form in a completely different manner than their small rocky cousins? As an alternative, Carnegie’s Alan Boss and others have suggested that gas giants may form rapidly through the direct collapse of disk fragments under their own gravity, rather than in a two-step process of first building up a rocky core and later gathering up a gas envelope.
Of course, astronomers do not have definitive answers to all the questions yet. But the results of laboratory experiments like Blum’s and theoretical calculations, which adhere to the laws of physics, confirm the basic storyline. Perhaps the most dramatic support comes from observational snapshots of different stages in the planetary birth process.
Snapshots in Time
As far back as the early 1980s, the Infrared Astronomy Satellite (IRAS) revealed that disks around young stars evolve over time. IRAS detected excess emission at mid-and far-infrared wavelengths, but not in the near-infrared, from several nearby “young-adult” stars, with ages of tens to hundreds of millions of years. The lack of near-infrared excess meant there wasn’t much hot dust very close to these stars, while the excess at longer wavelengths implied the presence of colder dust farther out. The simplest, and most compelling, explanation was that their disks had developed inner holes, possibly cleared out as a result of planet formation. What’s more, the disks were likely debris disks—that is, the dust comes from collisions of asteroids and evaporation of comets—and there is not much (if any) gas left.
Soon after the IRAS detections, Bradford Smith then at the University of Arizona and Richard Terrile at the Jet Propulsion Laboratory imaged a faint disk around one of these stars, named beta Pictoris, by blocking the light from the star itself with a coronographic mask. Their remarkable discovery provided our first glimpse of what might be a planetary system in the making. Not surprisingly, the beta Pictoris disk has been studied in every possible way with every new astronomical instrument ever since. Recent detailed images, from space and from the ground, show evidence of an inner hole and even a slight twist or two in the disk that might be caused by the gravity of planets embedded in it.
With the launch of Hubble and the advent of sensitive new infrared cameras on ground-based telescopes, astronomers were able to take pictures of dozens of protoplanetary disks in the 1990s. But just how long these disks live—thus the timescale for planet formation—remained uncertain. Besides, beta Pictoris remained the only disk imaged around a somewhat older star. That was the state of affairs when I flew down from Boston to Chile in March 1998 to use the 4-meter Blanco telescope at the Cerro Tololo observatory. Through a competitive proposal, I had been allocated four nights on it to collect the first dataset for my PhD thesis at Harvard. My goal was to look at a large sample of roughly 10-million-year-old stars to see what fraction of them still harbored dusty disks, thus to determine disk lifetimes. Most of the stars on my list were too far away to image the dust disks directly. Instead, I was looking for excess infrared emission that betrays a disk’s presence.
Located 2,500 meters above sea level, Cerro Tololo is a perfect place to do astronomy: clear, dry, and dark. Not that night. It was completely cloudy; we couldn’t even see the Moon, let alone the distant young stars I had come to investigate. Patricio Ugarte, the Chilean telescope operator, blamed the bad weather on El Niño. Being “clouded out” like this is an astronomer’s nightmare. There is not much you can do except sit and wait, hoping for the weather to improve, or perhaps watch a video or read a book and eat your “night lunch.” Charles Telesco and Scott Fisher were also in the control room with “Pato” and me. We controlled the telescope from there with a couple of computers. Charlie is an astronomy professor at the University of Florida, and Scott was one of his graduate students. Charlie’s team had built the electronic camera sensitive to the mid-infrared part of the spectrum (wavelengths roughly ten to twenty times longer than the reddest that the human eye can see). They had shipped the camera and its accessories, including cables and computers, in eight big crates all the way from Florida. Around 4 a.m., we gave up and decided to close down. There was no sign of the clouds’ parting.
We drove down to the dormitory area in the white VW Beetles that belonged to the observatory, and went to sleep. Not a happy way to start an observing run.
The next night, the clouds mostly cleared up. But it was still too humid; the water vapor in the air absorbs the infrared radiation from distant stars. I managed to observe a few of my target stars, but we needed much longer exposures because of the high humidity.
The third night was much better. Three more stars, nothing too exciting. Around 1 a.m., I decided to point the telescope at a closer and brighter star that was already known to have the signature of a dust disk. We took images with one filter. The star, designated as HR 4796A, appeared nice and bright. Then I decided to look at the star with a different filter, one that would capture longer wavelength radiation emitted by colder dust. Charlie was skeptical that it was worth the effort, given the high humidity, but I insisted. About twenty minutes into the exposure, we noticed something intriguing: the image building up on the computer screen appeared elongated rather than point-like. “Are we seeing the dust disk?” we wondered aloud. That seemed too good to be true. “Are you sure the telescope is not out of focus? Are you sure the focus doesn’t change when we change filters?” I asked. “No way, man. . . . At least, that’s never happened before,” Scott was the first to reply. Charlie, who had a lot more experience as an infrared observer, was also convinced this was the real thing. Once that exposure was over, we looked at a different star, just to make sure the focus was correct. It sure was. Now we knew: we had imaged the dust disk around the 10-million-year-old star HR 4796A. It was like looking at a baby solar system. We all cheered and high-fived each other; Charlie was almost dancing. “Congratulations, Ray! This is a big discovery. This is huge . . . ,” he beamed.
We continued to take more images of HR 4796A until dawn, just so that we had all the data necessary to convince other astronomers. By the time the Sun came up, we were tired yet feeling on top of the world. Scott and I stayed up another two hours to do a quick analysis of the images and sent them by e-mail to my PhD thesis advisors back in Boston. Once that was done, I tried to get a few hours of sleep, but that turned out to be impossible. I was way too excited. I couldn’t wait to talk to my advisors, but it was still too early in the morning for them to be in the office. After twisting and turning in bed for a few hours, I called my advisor Lee Hartmann. “It looks real to me,” he confirmed. I asked Lee to track down my other advisor Giovanni Fazio and convey the news to him.
Given that HR 4796A is at just the right age to be forming planets, we were intrigued to find evidence of a central hole about the size of the solar system in mid-infrared images of its nearly edge-on disk. The amount of dust in the disk adds up to only about an Earth mass: that is some thousands of times less material than what is found in 1-million-year-old Orion disks. Presumably, the original dust has gone somewhere, perhaps into building planetesimals. Of course, we do not see planets around HR 4796A, just circumstantial evidence in the form of an inner disk hole and a paucity of dust.
Soon after returning to Boston, I few to Gainesville to work with Scott and Charlie on further analysis of our images. It was while in Florida that we learned through the grapevine that another team of astronomers, led by David Koerner, then at the University of Pennsylvania, had captured images of the exact same disk the same week as we did, using one of the two Keck telescopes in Hawaii. It often happens in science that two or more independent groups hit upon the same quarry at roughly the same time. In this case, the apparent coincidence had a lot to do with the coming of age of mid-infrared cameras. We announced the discovery at a joint press conference at NASA Headquarters in Washington. It made news around the world, including the front pages of the New York Times and the Washington Post, because it was seen as a big step forward in tracing the origin of planetary systems.
The same week, a team of astronomers led by Wayne Holland and Jane Greaves, both then at the Joint Astronomy Center in Hawaii, presented millimeter-wave images of disks around four other, somewhat older, stars: Vega, Fomalhaut, epsilon Eridani, and beta Pictoris itself. The almost face-on disk around epsilon Eridani, which is a mere ten light-years away, is of special interest: it clearly shows a central cavity as well as a bright spot in the ring of dust. At an age of about 500 million years, that star must be well past the main epoch of planet formation. The dust ring is at roughly the same distance from epsilon Eridani as the Kuiper Belt of comets is from the Sun. Most likely, what we are looking at is the dust debris in a young Kuiper Belt analog around another star. The bright “blob” might be dust trapped in the orbit of an unseen planet. Newsweek magazine published a cover story, “The Birth of Planets,” in its May 4 issue, reporting on the HR 4796A disk as well as these four.
With adaptive optics in regular use on many of the largest ground-based telescopes, astronomers are now able to obtain images that are in some cases as sharp and sensitive as those from space-based observatories. Adaptive optics is a technique that partially corrects for the blurring effects of the Earth’s atmosphere; it works by flexing a thin mirror many times a second into just the right shape to cancel out the effects of roiling air above the telescope. In 2001, Kevin Luhman, then at the Harvard-Smithsonian Center for Astrophysics, and I imaged an edge-on disk around a T Tauri star in the MBM12 group using adaptive optics on the 8-meter Gemini North telescope. The disk appears as a dark lane in our images, with faint nebulosities on either side. The star itself is hidden behind the disk, but its light reflects off the top and bottom surfaces of the disk to produce the nebulosities. At a mere 2 million years, this star is still a toddler. Yet its disk is pretty thin—not as puffed up as the disks of million-year-old stars in the Taurus star-forming region. One possible explanation is that dust in the MBM12 disk has started to settle into the disk midplane, a bit like the dust particles that sedimented to the bottom of the jar in Blum’s experiment. If that is the case—something we still need to confirm—we may be seeing the first tentative steps toward planet formation already at the tender age of 2 million years.
Figure 2.3. Top: Our image of the disk around the young star HR 4796A. Bottom: A later image taken with the Hubble Space Telescope clearly shows it is shaped like a ring with a large cavity in the middle. The star itself has been masked out. Credits: R. Jayawardhana et al./NOAO (top) and G. Schneider (University of Arizona) et al./NASA (bottom)
Frozen Caches
Ralph Harvey describes himself as the “dog catcher of the planetary world.” That’s because he runs the Antarctic Search for Meteorites program, or ANSMET, which brings back hundreds of meteorites from the south polar region each year. It’s not that rocks from space arrive preferentially near the poles; they fall randomly all over the globe. But it’s easy to spot them against the Antarctic ice sheet, where there are virtually no other rocks on the surface. What’s more, the ice near Antarctic mountain ranges is exposed to fierce polar winds, so it sublimates, leaving the meteorites behind, which pile up over tens of thousands of years. The vast majority of them consist of extremely old, primitive materials—samples from the earliest days of the solar system. A small fraction comes from the Moon, Mars, or big asteroids that broke up in collisions.
As an eight-year-old, Harvey watched Neil Armstrong take those historic first steps on the Moon, and imagined himself as a space cadet with a jet pack and a ray gun. Having grown up in Wisconsin, he was used to snow and ice, but he wasn’t a mountaineer. At Beloit College, he did an undergraduate thesis on tektites, tiny glassy rocks thought to have formed during big meteorite impacts on Earth. A few years later, he entered graduate school at the University of Pittsburgh to work with William Cassidy, who ran ANSMET until Harvey took over the helm in 1991. Now a planetary scientist at Case Western Reserve University in Cleveland, Harvey had been to Antarctica twenty times by 2010.
It’s “the joy of being the first human being to see and touch a piece of space rock” that keeps him going back to one of the most alien environments on Earth. “Every time there’s a rock in front of me, something that nobody has ever seen before, it’s exhilarating. If that thrill wasn’t there, I wouldn’t be doing it. That’s what makes the hassles, and being away from family for a month or two over holidays, worthwhile.” He handpicks the roughly half-dozen expedition members each year, from among the hundreds that volunteer. Graduate students and scientists working on meteorites get preference. “People are often surprised by their own resilience, working in the extreme cold and isolation for up to two months” Harvey said. “Going to the toilet outside at 20 below zero for the first time is a memorable experience.”
According to him, with a little training, people are better at recognizing meteorites—“the one rock among many that just doesn’t look right”—than metal detectors or other instruments. Team members record the location using GPS devices and take photos before collecting each sample. At the end of the season, the entire collection is put in bags and sent by ship to the United States. Later the meteorites are catalogued and characterized by curators at the Johnson Space Center in Houston and the Smithsonian Institution in Washington. Finally, the catalog of new finds is published in the Antarctic Meteorite Newsletter, making it possible for researchers around the world to request specimens.
Most are run-of-the-mill stony meteorites known as chondrites, but a few dozen each year turn out to be unusually interesting. Harvey likens the status of meteoritic studies to where zoology was circa 1850. “The full range of samples is only now coming to light. We still need to fill in the empty niches,” he explained. “There may come a day when we say enough is enough. . . . We’re clearly not there yet.”
For example, two unusual meteorites recovered during the 2006 expedition consist mostly of the mineral feldspar, which is common in lunar rocks. Since it is relatively lightweight, feldspar is thought to have floated to the top of the magma ocean on the young Moon, forming a concentrated layer, while denser material settled in deeper. This process, called differentiation, would have occurred on other large bodies as well. The two meteorites, dubbed GRA 06128 and GRA 06129 after the Graves Nunataks ice fields where they were found together, contain a type of feldspar rich in sodium. Given their sodium feldspar concentration levels, scientists have concluded that the chunks came from a smashed-up dwarf planet. Collisions among such bodies must have been common during the planet-building epoch, and fragments like these give us glimpses of their long-vanished parents.
Even “ordinary” chondrites yield valuable clues as well as a long-standing mystery. They tell us a lot about the composition of the protoplanetary disk. Perhaps most important, the best measurement we have for the age of the solar system—4.566 billion years—comes from radioactive isotope dating of their constituents. What’s most puzzling about chondrites is their paradoxical mix of minerals that were once melted and others that clearly had a cold origin. In particular, millimeter-size pebbles embedded in them, known as chondrules, had to form at temperatures approaching 2000 degrees Celsius. Their roundness implies that the melting took place while the raw material was suspended in space, because that allows surface tension forces to pull them into spherical shape. But nobody is quite sure what caused the melting.
Frank Shu thinks the answer lies in his “x-wind” model for launching outflows in protostars. The idea is that a strong magnetic field of the star interacts with the inner edge of its disk to give rise to a wind, sort of like an eggbeater throwing egg from the center of the bowl to the outskirts. Shu and his colleagues suggest that x-winds from the proto-Sun lifted heated fluffy rocks and then sprayed them in a fiery rain all over the primitive inner solar system. These chondrules, or beads of melted rock, later combined with colder dust to form larger bodies like asteroids and planets. The scientists calculate that only those beads with sizes between a millimeter and a centimeter would fall back on the protoplanetary disk—in agreement with chondrule sizes seen in meteorites. What’s more, their model can explain the presence of unusual radioactive elements found in meteorites: fast-moving protons in energetic fares from the young Sun would combine with ordinary elements in the protoplanetary disk to form radioactive counterparts. But meteorite researchers point out a number of issues with Shu’s theory. For example, some chondrules appear to have been heated more than once. It’s not clear how origin in a stellar wind could account for such multiple heating events.
Other proposals include lightning in the solar nebula, collisions between molten planetesimals, and shock waves propagating through the protoplanetary disk. The latter is the current favorite. Recent calculations suggest that shocks can flash-heat particles in the disk to about 1800 degrees Celsius for a few minutes, fusing them into chondrules, which then take several hours to cool down. That sequence of events and the timescales appear to agree with what scientists infer about the thermal history of these beads. What generated the waves in the first place is still being debated. Astronomers’ best guess is that building blocks of Jupiter spawned spiral waves in the protoplanetary disk, and those waves piled up into shock fronts in the inner part of the disk, like breakers hitting a beach.
Slingshot Games
Taken together, the evidence to date suggests that dusty disks evolve into infant planetary systems within about 10 million years. But that is by no means the end of the action. It could take hundreds of millions of years before planetary systems achieve their mature form. In our own solar system, 50 million years after the Sun’s birth, chunks of rock that hadn’t made it into agglomerating planets were still flying about in chaotic ways. Not all the newborn planets had settled into their uneventful dance around the Sun yet. One of them, about the size of Mars, was actually on a collision course with the Earth. It was perhaps half the Earth’s diameter and one-tenth the mass. This roaming planet hit the still-warm Earth at some 40,000 kilometers per hour, sending a huge plume of material into space. In the throes of the collision, Earth’s primordial atmosphere boiled off into space and its mantle melted into an ocean of magma. Out of the debris of that catastrophic event, scientists now believe, the Moon was born.
Scientists have mulled over different ways of making the Moon for well over a century. George Darwin, son of the famous naturalist Charles Darwin, was among the first to put forth a model for the Moon’s origin. In 1878, he suggested that a newly born, still molten Earth started spinning faster and faster until it threw off a piece of itself as big as the Moon, sort of like a merry-go-round spinning out of control and sending a kid flying. But Darwin’s model fails an important physical test: it can’t explain the total spin rate—a quantity that physicists call angular momentum—of the Earth-Moon system. If his theory were correct, both bodies would spin much faster than they actually do. A second theory suggested that the Moon assembled itself, independent of the Earth, from primitive rocks and dust, just as other planets in the solar system did. In that case, both bodies should have similar percentages of iron. But the Moon’s core has far less iron. A third possibility is that it had formed elsewhere in the solar system and was later captured by the Earth’s gravity. Theoretical calculations show that capture into orbit is highly unlikely: if a Moon-size body were to come near the Earth, it’s a lot more likely to have either hit the Earth directly or received a gravity kick that set it flying off into space.
In the early 1970s, two sets of theorists—one consisting of Alistair Cameron and William Ward at Harvard and the other of William Hartmann and Donald Davis of the Planetary Science Institute in Tucson—independently suggested that the Moon formed from the debris of a giant impact that the Earth had with a Mars-size roaming planet. But it took nearly a decade before planetary scientists widely accepted that catastrophic impacts had been common in the early solar system. In computer simulations of the event, the impactor is destroyed, and a plume of rock, magma, and vapor is boosted into Earth orbit. Occasionally, a fairly large rocky body is formed, in addition to the small debris. The impactor’s iron core falls onto the deformed proto-Earth and sinks to its center. That explains why the Moon has very little iron; after all, the debris that went into it came from the rocky mantle material of the impactor and the Earth. The model can also account for the paucity of water and other volatiles on the Moon, because volatiles in the hot debris would have escaped into space. Radiometric dating of lunar rocks reveals that the Moon was assembled 50 million years after the solar system itself. One potential stumbling block for the giant impact theory remains: the energetic event should have melted the Earth’s mantle, altering its composition with different minerals rising to the top or sinking to the bottom. Geochemists have found no signs of such alterations. It’s possible that the evidence was wiped out by 4 billion years of geologic activity.
By now, most researchers accept that collisions among hundreds of planetesimals—some as big as the Moon—were required during the first 100 million years to build up the planets to their present masses. What’s perhaps more puzzling is the evidence for an intense period of planetesimal bombardment some 600 million years later. The cratering record on the Moon shows a sharp peak at 3.9 billion years ago, as if the inner solar system were pelted by a furry of large meteorites for a brief period long after the planets formed. Life on Earth, if it had developed by then, would have been disrupted or even reset. The cause of this “late heavy bombardment” remains a mystery. Some have suggested that the growth and outward migration of Neptune may have catapulted millions of leftover planetesimals to the inner solar system. Others blame the possible inward migration of Jupiter, which may have stirred up the asteroid belt, especially if Jupiter and Saturn entered into resonant orbits (such that Jupiter would circle the Sun twice as Saturn goes around once). We do not know for sure whether that actually happened. New clues about our solar system’s adolescence are likely to come from observations of planets orbiting other stars.