A reprint from American Scientist the magazine of Sigma Xi, The Scientific Research Society

This reprint is provided for personal and noncommercial use. For any other use, please send a request to Permissions, American Scientist, P.O. Box 13975, Research Triangle Park, NC, 27709, U.S.A., or by electronic mail to [email protected]. ©Sigma Xi, The Scientific Research Society and other rightsholders

“Stellar embryology” takes a step forward with the first detailed look at the youngest -like

Thomas P. Greene

natomists have been studying the of stellar birth is more than 400 light- in the past decade, astronomers have Aembryonic development of ani- years away. At such distances, a prena- now been able to observe these seclud- mals for centuries. Their detailed de- tal the size of our Sun is exceed- ed stellar embryos. My colleagues and I scriptions of tissue growth may have ingly dim. have been among the privileged few to reached a high art, but a theoretical un- Despite such impediments, astro- witness some of the earliest stirrings of derstanding of how the embryo under- physicists have been able to piece to- a star’s life. goes its remarkable transformation gether the broad outlines of how a low- from a single cell to a crying infant lags mass (Sun-like) star forms. It’s a Baby Steps to Stardom far behind. Until recently, the field of surprisingly complex process. Al- The process of making a star begins in- astrophysics had been struggling with though it may involve the simplest of side enormous clouds of interstellar a nearly opposite problem. Theoretical elements and molecules, the making of gas and dust that lie primarily within models of how a star forms—its em- a star is directed by a maelstrom of the plane of the (Figure 1). Con- bryonic development from a cold competing forces—including gravita- sisting chiefly of molecular hydrogen cloud of interstellar gas to a blazing tional collapse, magnetic fields, nuclear (H2), the largest of these giant molecular furnace of nuclear fusion—had out- processes, thermal pressures and fierce clouds may be several hundred light- paced the astronomer’s ability to ob- stellar winds—all of which wish to years across and weigh as much as serve the process. It’s as if develop- have their way with the unformed star. 100,000 solar masses. These clouds mental biologists had identified the Because the interaction of these forces have an average density of about 100 major stages in the growth of an em- is not fully understood, there is much H2 molecules per cubic centimeter, but bryo with only the crudest views of that remains mysterious about the they are far from being homogeneous. their subject. birth of a star. How exactly does a Some regions may have a density of This peculiar state of affairs in astro- growing star accrete matter from its physics is largely due to the reclusive surroundings, and how does the nature of embryonic stars. A stellar process stop? Why do stars form in the embryo grows in a cloudy womb of numbers and range of sizes that they molecular gas, which is so choked with do? And why do some stars form plan- dust that visible light has little hope of etary systems? These are fundamental passing through. A view of these mole- questions that cannot be answered cular clouds in even the largest optical without actually observing the process telescope would reveal little more than of . a dark patch of sky. Observations are Fortunately, the field of stellar em- further hindered by the forbidding dis- bryology has recently turned a corner tances to the stars. The nearest region as an observational science. Although visible light is unable to penetrate the Thomas Greene is an astrophysicist at NASA’s dusty cloud that swaddles a prenatal Ames Research Center near Mountain View, star, the longer wavelengths of California. He has also held positions at the radiation can easily slip through the Lockheed Martin Advanced Technology Center dust and so escape the inner confines and at the University of Hawaii Institute for of the cloud. Such radiation has been . In addition to his work on young detected for decades, but until recently stars, he has also done research in developing infrared telescopes have lacked the sen- infrared detectors and various astronomical instruments. In his spare time he bicycles and sitivity and resolution to provide de- hikes with his wife Deb and their dog Sputnik. tailed information about the youngest Figure 1. Stellar embryos—protostars—lie Address: NASA Ames Research Center, MS prenatal stars, the protostars. With the hidden within the dust of the Rho Ophiuchi 245–6, Moffett Field, CA 94035–1000. Internet: development of giant telescopes and dark cloud, a star-forming region about 400 [email protected] extremely sensitive infrared detectors light-years away, near the common border of

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 316 American Scientist, Volume 89 with permission only. Contact [email protected]. Anglo-Australian Observatory/Royal Observatory, Edinburgh the Ophiuchus and Scorpius. Optical-wavelength images, such as this one, cannot reveal these young stellar objects, but a new generation of large telescopes and sensitive infrared detectors has allowed astronomers to make detailed observations of these objects for the first time. Light from Rho Ophiuchi (near top)—a hot young star associated with the dark and dusty cloud—is reflected as a bluish glow. The aging red- Antares (lower left) is not related to the star-forming region.

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 2001 July–August 317 with permission only. Contact [email protected]. a dark cloud b gravitational collapse c envelope bipolar flow

disk

dense core

10,000 to 200,000 AU 10,000 AU time = 0 500 AU 100,000 years d e pre-main-sequence star f young stellar system bipolar flow planetary debris central disk star

central planetary

100,000 to 3,000,000 to after 100 AU 3,000,000 years 100 AU 50,000,000 years 50 AU 50,000,000 years

Figure 2. Early development of a young, Sun-like star can be described in a series of stages that span more than 50 million years. Star for- mation begins inside dark interstellar clouds containing high-density regions (a) that become gravitationally unstable and collapse under their own weight (b). The collapsing core forms a protostar (c), a phase of defined by the rapid accumulation of mass from a circumstellar disk and a surrounding envelope of gas and dust. As the dusty envelope dissipates, the object becomes visible at optical wavelengths for the first time as a T Tauri star (d). These objects can often be recognized in telescopic images by the presence of a protoplanetary disk (see Figure 5). After a few million years the dusty disk dissipates, leaving a bare pre-main-sequence star at its center (e). In some instances, a debris disk with newly formed may continue to orbit the star. The star continues its gravitational collapse to the point where its core temperature becomes hot enough for nuclear fusion, and the object becomes a main-sequence star (f). (AU = astronomical unit, the average distance between the Sun and the Earth.)

10,000 molecules or more per cubic about 100,000 years) is defined by the which is the product of rotational ve- centimeter, or more than 1,000 times rapid accumulation of mass from the locity and radius—of the core material the density of the most rarefied parts of surrounding envelope of gas and dust remains constant, so material rotates the cloud. A may have (Figure 3). It does so at the rate of a few faster as it gets closer to the protostar. many such dense cores (each of which millionths of a solar mass—equivalent Slowly moving material falls directly may become a star), but the star-forma- to one Earth-sized —every year. onto the protostar, but some of the gas tion process is not very efficient, as a Coinciding with the onset of and dust is moving so quickly that it 100,000-solar-mass cloud never yields is a steady outflow of material in pow- travels in an orbit instead. Since all of 100,000 Sun-sized stars. In fact, the con- erful winds that emanate from the the material in the envelope surround- version efficiency of cloud mass to stel- poles of the young star. These bipolar ing the protostar rotates in the same di- lar mass probably averages less than jets are telltale signs of a protostellar rection, the matter falls into orbits of 10 percent. system, and it is ironic that protostars various sizes depending on its velocity, Giant molecular clouds are support- are often detected by the jets that shed and so forms a circumstellar disk. Most ed against their weight by thermal mass from the system, rather than the matter eventually flows onto the proto- pressure, turbulent gas motions and process of accretion (Figure 4). star through the disk, but some of it re- magnetic fields within, but at some Throughout this period, the proto- mains in orbit. As the surrounding en- point their dense cores become gravita- star progressively increases in density velope of dust disperses, the accretion tionally unstable and begin to collapse as it shrinks in size. The infalling mate- process stops, and the central globe of (Figure 2). The center of the collapsing rial, which had been rotating relatively gas is no longer considered to be a pro- core becomes a protostar, which, as the slowly around the dense core, begins to tostar; it is now a pre-main-sequence (or name suggests, is the first step toward speed up as the radius of the protostar PMS) star. (Protostars and PMS stars stardom. The protostar’s life (lasting decreases. The angular momentum— are often grouped under the term young

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 318 American Scientist, Volume 89 with permission only. Contact [email protected]. stellar objects, which neatly includes all of the gas, which acts to expand the stellar evolution is extremely stable, prenatal stars.) star. The gas pressure is proportional and may last for many billions of The pre-main-sequence phase of evo- to the star’s temperature, which is only years. Our Sun is a fine example of a lution lasts for tens of millions of years. about 3,000 to 4,000 kelvins (degrees main-sequence star—one that has been In its earliest phases, the first few mil- above absolute zero) in the outer re- steadily burning hydrogen for 5 billion lion years or so, these objects are often gions but nearly one million kelvins in years, and should continue to do so for called T Tauri stars, a name derived from its core. This represents a dramatic in- another 5 billion years. The Sun’s rela- the archetypical star in the crease from the frigid 10 to 20 kelvins tive stability is the key reason that crea- Taurus. Having shed their dusty en- typical of the dense cores in a molecu- tures such as ourselves can evolve on velopes, T Tauri stars are the youngest lar cloud, but one million degrees is the “debris” that continues to orbit the objects that can be seen with an optical still considered to be “cool” for a stellar star after it is formed. telescope. They are still surrounded by a core. This temperature is just hot disk of dust and gas—often called a enough to fuse a deuterium atom (hy- Youthful Energy protoplanetary disk at this stage—and drogen-2) and a proton into helium-3, a Our Galaxy appears to have a healthy may continue to eject material in their process that releases only a modest lust for making stars, with many places bipolar jets. After a few million years, amount of energy. that can be aptly called “stellar nurs- much of the dust and gas in the proto- It takes tens of millions of years, but eries.” Some of these nurseries may planetary disk dissipates, leaving a bare eventually the crushing force of consist of merely a few young stars, PMS star in the center. In some in- wins the battle against the outward whereas others may contain many stances, a few large bodies may continue thermal pressure of the gas within the hundreds. Among the more notable of to orbit the star in a remnant debris disk star. The compression of material raises these are the Orion Nebula and the (Figure 5). The planets, moons and aster- the temperature of the star’s interior to star-forming clouds near the star Rho oids of our all had their be- about 10 million kelvins—hot enough Ophiuchi. These regions have been ginnings in such a disk. to fuse four protons into one helium-4 studied for decades and have provided At this stage the internal structure of atom, which is the primary generator the observational basis for much of the star is determined by a balancing of energy within a true star. The event what we do know about the first stages act between gravity, which compresses marks its arrival on the main-sequence of star formation. and heats the object, and the pressure phase of its evolution. This period of The PMS stars were among the first

dusty envelope

protostar

circumstellar disk Courtesy of the European Southern Observatory Courtesy of the European

Figure 3. Protostars accumulate mass from a surrounding envelope of dust by way of a cir- Figure 4. Herbig-Haro 34, a young stellar cumstellar disk that orbits the young object. Most of the energy the protostar radiates comes object, exposes one of its bipolar jets (red from this accretion process, which typically ends after about 100,000 years, as the object streak), which emanates from the central approaches its final mass. Powerful winds, which appear to be linked to the accretion star. Two bow-shaped regions (upper right process, emanate from the poles of the star, expelling some of the envelope’s mass. These and lower left) show where the jets collide bipolar jets are often evident in telescopic images (see Figure 4). with the .

© 2001 Sigma Xi, The Scientific Research Society. Reproduc- 2001 July–August 319 tion with permission only. Contact [email protected]. a c

b

Figure 5. Protoplanetary disks (or ), seen on edge (a) or face on (b), can be found in images of star-forming regions. These dust-filled disks may evolve to form planetary systems, but some recent images of the trapezium in Orion (c) suggest that strong winds from energetic neighboring stars may destroy the proplyds (yellow tear drops pointing in the direction of the wind) before the planetary systems form. (A and B courtesy of Mark McCaughrean, C. Robert O’Dell and NASA. C courtesy of John Bally, David Devine, Ralph Sutherland and NASA.) young stellar objects to be observed, main-sequence counterparts of the was still in the process of contracting. which is not surprising given that same temperature. Since an object’s lu- Protostars were not discovered until many of them are optically visible. In minosity is proportional to its radius the pioneering infrared surveys of the the 1950s, Merle Walker, then at the (R) and temperature (T), in the relation 1960s and 1970s, which found several Mount Wilson and Palomar Observa- R2T4, it was apparent that these objects candidates embedded deep within the tories, noticed that some stars in the were larger than the main-sequence dark clouds. Space-based observations vicinity of dark clouds were intrinsical- stars. This is precisely what theoreti- later showed that protostars are up to ly brighter (more luminous) than their cians had predicted for a PMS star that 10 times more luminous than PMS

protostar envelope PMS star with circumstellar disk PMS star without disk (30 kelvins) disk (many temperatures)

scale) central star × (4,000 kelvins) central star central star relative flux relative flux relative relative flux (10 relative

2.2 10 100 1,000 2.210 100 1,000 2.2 10 100 1,000

wavelength (micrometers, log scale) Figure 6. Spectral energy distributions at infrared wavelengths can reveal different classes of young stellar objects. The key to the identifica- tion is the relation between wavelength and temperature, with longer wavelengths corresponding to cooler temperatures. Protostars (left) emit a broad band of infrared energy, with a peak near 100 micrometers, which emanates from the cold envelope of material surrounding the star. A circumstellar disk is recognized as a broad range of emissions that decrease in strength at cooler temperatures (middle), corresponding to a range of distances from the central pre-main-sequence (PMS) star. A bare PMS star is recognized by a relatively narrow peak near 2 micrometers (right), which corresponds to a temperature of about 4,000 kelvins. (Adapted from the work of Charles Lada and others.)

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 320 American Scientist, Volume 89 with permission only. Contact [email protected]. stars. But since protostars are not much bigger than PMS stars, it suggested that they had an extra source of energy that was contributing to the protostar’s . In the decades that followed, as- tronomers refined their observations of young stellar objects. Such studies in- cluded the object’s spectral energy distrib- ution—the amount of energy the object releases over a range of wavelengths (especially in the infrared part of the spectrum). As it happens, measure- ments of an object’s spectral energy dis- tribution are surprisingly informative. In particular, the shape of the dis- tribution reveals the evolutionary stage of the object (Figure 6). Since the various components of the object—the central star, the dusty envelope and the circumstellar disk—each have dif- ferent temperatures, the amount of in- frared radiation the object emits at a particular wavelength tells us which components are present. The spectral energy distribution of a protostar, for example, is dominated by an infrared Figure 7. Highly sensitive infrared spectrographs, such as the SpeX (blue box) and CSHELL emission near 100 micrometers, which (red box) instruments at NASA’s Infrared Telescope Facility on Mauna Kea, allow represents the radiation from the large astronomers to measure the atomic spectra of protostars for the first time. Here the spectro- outer reaches of its chilly (30-kelvin) graphs are attached to the bottom of the 3-meter IRTF reflector telescope. (Image courtesy envelope. The presence of the enve- of the IRTF staff.) lope also explains why protostars are more luminous than PMS stars: Mat- as provide an accurate reading of its clouds. These turn out to be PMS stars, ter from the envelope is still falling temperature, its radius and its rate of which are very similar to their optical onto the protostar. As the material rotation. counterparts, except that they are still slams into the protostar’s surface, its Modern telescopes and infrared de- hidden in the dust. However, there are a gravitational energy is converted to tectors have been able to show the ab- number of young stars in the dark thermal energy—the source of the in- sorption lines in the spectra of some clouds that do not show strong spectral frared radiation. young stars still embedded in dark absorption features. The spectral energy In contrast, the spectral energy dis- tribution of a PMS star with a circum- stellar disk is dominated by the central star, with a peak emission near 2 mi- star crometers. The disk itself shows up as a broad range of emissions, especially at longer wavelengths, representing the various temperatures of the disk (which decrease with distance from the protostar central star). In the absence of the dusty envelope, the accretionary process has slowed to a trickle, if anything at all, in relative flux the PMS star. As valuable as these measurements have been to our understanding of star formation, they only scratch the sur- Sc Na Si Sc Fe Na Vn face when it comes to revealing the physical properties of the young stars 2.206 2.207 2.208 2.209 2.210 themselves. A much better probe is a wavelength (micrometers) star’s true spectrum, which shows atomic and molecular absorption lines Figure 8. Atomic spectrum of a protostar can be matched to a “normal” star such as a red (Figure 8). Such spectra reveal the iden- dwarf, a relatively small, cool star. The absorption line in a spectrum reveals the elements tity of the elements and molecules pre- and molecules in the protostar’s atmosphere, and can be used to deduce the star’s tempera- sent in the star’s atmosphere, as well ture, radius and rate of rotation. The data were collected with CSHELL (see Figure 7).

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 2001 July–August 321 with permission only. Contact [email protected]. 10 scopes collect infrared radiation with R their large primary mirrors, which fo- h o cus the light into spectrographs. These O p spectrographs use large diffraction 1.5 M h i gratings to disperse the light (into its u c h component wavelengths), which is i p then focused again onto sophisticated r o infrared detector arrays (Figure 7). The t o s optics and the detectors must be cooled t a to as low as 30 kelvins (with liquid ni- r s trogen and mechanical coolers) to pre- 1.0 vent the infrared radiation of the in- struments from overpowering the faint 1.0 M signals from the young stars. 100,000 Charles Lada, of the Smithsonian years old Astrophysical Observatory, and I have (birthline) recently used the IRTF and Keck tele- scopes to make the first detections of 0.8 M infrared absorption lines in Sun-like protostars. The task required a large 500,000 amount of observing time (about a years old week per year) to build up a sufficient 0.1 1,000,000 signal. So far we have recorded about a years old dozen high-resolution spectra of

protostar luminosity/sun’s luminosity (log scale) protostar luminosity/sun’s 0.6 M strongly veiled protostars in the dark clouds near Rho Ophiuchi. Interestingly, the atomic spectra of these objects are similar to main-se- 3,000,000 quence stars that are very cool—except years old that their absorption features are much 0.4 M weaker. (Main-sequence stars such as M = solar mass red dwarfs can have very cool surfaces, 0.2 M even though their core temperatures 0.01 are hot enough for thermonuclear fu- 8,000 6,300 5,000 4,000 3,100 2,500 sion.) We interpret this to mean that the protostars have surface temperatures surface temperature (kelvins, log scale) that are similar to the cool stars, about Figure 9. Hertzsprung-Russell diagram reveals the evolutionary course of a star based on its 3,500 to 4,000 kelvins. This is about the mass (blue lines). A cohort of stars that were born together move as a wave through time same as the PMS stars in the same (green lines). The protostars of the Rho Ophiuchi dark cloud define the birthline for low- clouds, which have temperatures as mass stars, and thus represent the youngest stars ever plotted on such a graph. Here only the high as 4,500 kelvins. We need more ob- first few million years of evolution are shown, even though most low-mass stars have life- servations to see whether there are tru- times measured in billions of years. The early evolution of our Sun is represented by a ly any significant differences in the tem- of 1.0 solar mass. peratures or sizes of these young stars. We have also quantified the veiling distributions of these objects suggest and the energy liberated by matter of these protostars by measuring the that they are protostars, and they offer a falling in from the envelope and disk. depth of their absorption lines relative special challenge for stellar spectroscopy. The glow of the warm dust usually to main-sequence stars and PMS stars emits several times as much near-in- of equivalent temperatures. At wave- Playing Peek-a-Boo with Protostars frared radiation as the central star it- lengths of about 2 micrometers, some It’s been said that protostars are the self. The protostar is effectively washed of the veiling is as much as three times “Holy Grail” of infrared astronomy, out—or veiled—by its surroundings. brighter than the star itself! We think and this statement pretty much sums What is an infrared astronomer to do? this “excess” emission is caused by the up the difficulty of the pursuit. The The solution lies in bigger telescopes active accretion of material either onto dust that surrounds a protostar dims and extremely sensitive infrared detec- the protostar or within its circumstellar its visible light by a factor of one bil- tors. Fortunately, such equipment now disk. In some instances the veiling is lion, and even its near-infrared emis- exists. Three of the more powerful in- even greater than a threefold bright- sions (wavelengths around 2 microme- frared telescopes are located atop Mau- ness of the central star. We interpret ters) are dimmed by a factor of 10. na Kea on the big island of Hawaii: the this higher level of veiling as the accre- Much of this dust is very close to the 3-meter NASA Infrared Telescope Fa- tion of matter from the circumstellar star (about 5 stellar diameters away), cility (IRTF), the 3.8-meter United King- envelope. and is heated to a temperature of near- dom Infrared Telescope and the giant By subtracting the excess emissions ly 1,500 kelvins by the star’s radiation 10-meter Keck II telescope. These tele- we can also determine how much in-

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 322 American Scientist, Volume 89 with permission only. Contact [email protected]. frared radiation comes from the central star itself. This allows us to estimate the protostar’s intrinsic brightness, which reveals a luminosity comparable to the PMS stars in the same cloud. This sug- 40 kilometers gests that protostars and PMS stars are per second about the same size. These measures of the protostar’s luminosity may appear to be in slight conflict with their spec- tral energy distributions, which indi- cate that protostars are several times 30 kilometers brighter than PMS stars. However, the per second spectral energy distributions of the pro- tostars also include the accretion of matter onto their central stars, a process that releases gravitational energy. The temperatures and 20 kilometers per second of the protostars can also be converted flux relative to estimates of their ages and their masses. Theorists generally model the evolution of PMS stars, but are now be- ginning to consider the effects of accre- tion so that their models can be applied 10 kilometers to protostars. This was recently done per second by several independent groups of astrophysicists, including three Italian theorists, Francesco Palla, Francesca D’Antona and Italo Mazzitelli, as well as Steven Stahler of the University of protostar California, Berkeley, and Isabelle (about 25 kilometers Baraffe and Lionel Siess of France. per second) Their sophisticated calculations allow us to place the protostars on a temper- ature-luminosity—or Hertzsprung- Russell (H-R)—diagram for the first 2.293 2.294 2.295 time ever (Figure 9). This is no small ac- wavelength (micrometers) complishment, since the H-R diagram is the key to providing the age and Figure 10. can be measured by the absorption-line spectrum of the carbon mass of a star. monoxide molecule in a protostar’s atmosphere. A computer compares a protostar’s noisy All of the protostars in the plotted spectrum (red line) to the shapes of synthetic spectra (yellow lines) from objects with differ- sample come from the same dark ent rotation velocities. cloud near Rho Ophiuchi. They ap- pear to be very young, effectively allow us to measure is the star’s rate of question is to observe how the rate of defining the observational birthline rotation. All stars spin on their axes, a rotation changes during the early where stars first appear on the H-R di- quality that is believed to be inherited stages of a star’s formation. agram, and all appear to be light- from the rotation of the dense core as it The key to measuring a protostar’s weights of about 0.5 solar mass. Simi- collapses in the prenatal dark cloud. rotation is the carbon monoxide mole- larly, PMS stars in the same cloud Our Sun has a rotation period of about cule. At a wavelength of about 2.3 mi- appear to span a range of masses with 26 days, which corresponds to a veloc- crometers, the absorption line of carbon a peak near 0.5 solar mass. The masses ity of about 2 kilometers per second at monoxide has detailed structure that is of protostars and PMS stars are gener- its equator. This actually corresponds very sensitive to the rotation of the pro- ally expected to be similar because the to a tiny fraction (one percent) of the tostar, but is not altered by the star’s protostars should have accreted nearly angular momentum in our solar sys- other physical properties (Figure 10). By all of their final mass at this stage of tem, even though the Sun contains the comparing the observed shapes of this their evolution. However, many more bulk of the system’s mass. Instead, feature to synthetic spectra with differ- protostars will have to be studied to most of the solar system’s angular mo- ent rotation velocities we can determine determine whether there is a signifi- mentum lies in the planets, especially the object’s rate of rotation. Actually, cant difference between their masses Jupiter and Saturn. One of the myster- what we do measure is the object’s pro- and those of PMS stars. ies of solar-system formation is how jected rotation velocity. Since we don’t the bulk of the mass gets concentrated know the exact alignment of the object’s Spinning in the Cradle in the center, while almost all of the an- axis to our line of sight, we are not get- Among the more interesting physical gular momentum is transferred to the ting a full measure of the object’s rate of properties that absorption-line spectra periphery. One road to answering this rotation. Even so, this information is ex-

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 2001 July–August 323 with permission only. Contact [email protected]. pre-main-sequence protostar 40 kilometers per second star 20 kilometers per second magne etic c c co u p li in g gnetic g a c m o u p l i i n n

g

g

Figure 11. Protostars rotate more rapidly than pre-main-sequence stars, suggesting that young stellar objects slow down as they age. One the- ory holds that protostars and pre-main-sequence stars are coupled to different parts of the disks that surround them. Since the inner region of a circumstellar disk rotates more quickly than the outer region, theorists argue that protostars may be closely coupled to the inner parts of their disks, perhaps because they are still accreting matter from the disks. If so, then the inner edge of a protostar’s disk would also be about four times closer to its central star than is the disk of a pre-main-sequence star. tremely valuable because the orienta- slowing the spinning star. The phenom- arise from the stars’ atmospheres and tion of these objects should be more or enon can be likened to pulling on a not their circumstellar disks. However, less random with respect to the Earth. sticky wad of bubble gum that is at- we have found a handful of young The results are intriguing. Nearly all tached to a spinning baseball. The ball stars that intermittently show disk-like of the protostars that we observed are may not stop spinning, but it will slow features. They turn out to be FU Orion- rotating quickly, with projected veloci- down. Whatever the mechanism, our is variables, which also show large, ties greater than 20 kilometers per sec- results suggest that the transition from episodic changes in their luminosity. ond. On the other hand, most PMS protostar to PMS star involves a drop in One possible explanation for this puz- stars with circumstellar disks are rotat- rotational velocity and a coincident zling observation is that the FU Orion- ing slowly, with projected velocities drop in angular momentum. is protostars accrete material through less than 20 kilometers per second. The absorption-line spectra also re- their disks in brief spurts. Most of the What accounts for the differences in ro- veal some other interesting phenome- time they would be in low-accretion tation velocities? na involving the accretion process. states, during which they would show Several theorists have shown that When we look at an object’s spectrum star-like spectral features. Occasionally, the rotation velocity of a young star we must distinguish whether it arises however, they would undergo periods can be controlled by its circumstellar from the atmosphere of the central star of high accretion, when their luminosi- disk through the star’s magnetic field. or from its circumstellar disk. Since the ty increases and their spectra become If this process occurs in both types of accretion process will make the disk dominated by disk-like features. This young stellar object, it means that the hot and relatively dense, it should have idea, largely championed by Lee Hart- protostars must be coupled to regions its own emission spectrum and ab- mann of the Smithsonian Astrophysi- of their disks that are spinning faster sorption lines. Fortunately, the physi- cal Observatory, is currently the best than those of PMS stars (Figure 11). cal differences between the disk and explanation out there. Since Kepler’s third law dictates that the star help us to discriminate be- Might all protostars go through an the circumstellar material that orbits a tween their spectra. The disk is less FU Orionis phase? So far the infrared star most quickly must be closest to the dense than the star, so its gas is under spectra of protostars and hundreds of star, our results suggest that protostars less pressure, which means that certain PMS stars show little or no evidence of are coupled to nearby parts of their elements will have different levels of FU Orionis-like features. If the FU Orio- disks. This may arise because proto- ionization (which are apparent in the nis phase of evolution were common to stars are still accreting large amounts spectrum). Moreover, disks rotate at all young stars, one would expect at of matter directly from their disks. Or it different speeds at each radial distance, least a few of these stars to show some may be that the accreting matter car- whereas stars rotate essentially as a sol- evidence of accretionary spurts. Perhaps ries a substantial angular momentum, id body. These physical differences the FU Orionis stars are merely freaks. which “spins up” the protostars. translate into substantial differences in In contrast, PMS stars are no longer the relative strengths of the absorption Conclusion accreting much matter and instead may lines and the velocity profiles of the Many of the protostars that we’ve ob- have experienced some braking. “Disk carbon monoxide features. served in the infrared won’t be seen in braking” is thought to be produced by The spectral features of nearly all visible light for another hundred-thou- ionized gas in the circumstellar disk protostars that my colleagues and I sand years or more. Yet we now have a which tugs on the star’s magnetic field, have studied clearly indicate that they fair gauge of the qualities that as-

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 324 American Scientist, Volume 89 with permission only. Contact [email protected]. tronomers seek to measure for true Greene, T. P., and C. J. Lada. 1997. Near-in- Meyer, M. R., F. C. Adams, L. A. Hillenbrand, J. stars: their luminosities, sizes, ages, frared spectra of flat-spectrum protostars: M. Carpenter and R. B. Larson. 2000. The extremely young revealed. stellar : constraints masses, rates of rotation and the pres- The Astronomical Journal 450:233–244. from young clusters and theoretical per- ence of circumstellar matter. It is merely Haisch, K. E., Jr., E. A. Lada and C. J. Lada. spectives. In Protostars and Planets IV, ed. V. a beginning. There is much that remains 2000. A near-infrared L-band survey of the Mannings, A. Boss and S. Russell. Universi- unanswered, and many more dark young embedded cluster NGC 2024. The As- ty of Arizona Press. clouds must be probed for protostars. tronomical Journal 120:1396–1409. O’Dell, C. R., and S. V. W. Beckwith. 1997. Young stars and their surroundings. Science Nevertheless, even the early anatomists Hartmann, L. 1998. Accretion Processes in Star 276:1355–1359. Formation. Cambridge University Press. began their studies of embryonic devel- Shu, F. H., F. C. Adams and S. Lizano. 1987. opment with modest descriptions. Per- Lada, C., and N. Kylafis. 1999. The Origins of Star formation in molecular clouds: obser- haps someday the science of stellar em- Stars and Planetary Systems. Proceedings of vations and theory. Annual Reviews of As- bryology may too describe the process the NATO Advanced Science Institutes, tronomy and Astrophysics 25:23–81. Crete 1998. New York: Kluwer Academic of star formation in fine detail. Stahler, S W. 1994. Early stellar evolution. Pub- Publishers. lications of the Astronomical Society of the Pa- Acknowledgment Lada, C. J. 1987. Star formation: from OB asso- cific 106:337–343. The author wishes to thank Charles Lada, ciations to protostars. In Star Forming Re- who has worked with him on many of the gions, ed. M. Peimbert and J. Jugaku. Pro- ideas and results in this article. ceedings of the Symposium, Tokyo, Japan, Nov. 11–15, 1985 (A87–45601 20–90). Dor- Bibliography drecht, Netherlands: D. Reidel Publishing Co., p. 1–17; discussion, p. 17, 18. André, P. 1995. Low-mass protostars and pro- tostellar stages. Astrophysics and Space Sci- Lada, C. J., E. T. Young and T. P. Greene. 1993. ence 224:29–42. Infrared images of the young cluster NGC Greene, T. P., and M. R. Meyer. 1995. An in- 2264. The Astrophysical Journal 408:471–483. frared spectroscopic survey of the rho Ophi- Luhman, K. L., and G. H. Rieke. 1999. Low- uchi young stellar cluster: masses and ages mass star formation and the initial mass from the H-R diagram. The Astrophysical function in the rho Ophiuchi cloud core. The Journal 450:233–244. Astrophysical Journal 525:440–465.

© 2001 Sigma Xi, The Scientific Research Society. Reproduction 2001 July–August 325 with permission only. Contact [email protected].