A minor revolution in astronomy occurred on April 6, 1992. It did not take place at a mountaintop observatory but happened at anunlikely location the Callaway Gardens Inn on Georgia's Pine Mountain (elevation: 820 feet). Astronomers had gathered therefor an international meeting on the normally slow-paced researchtopic of double stars, a field where discoveries often require decades to allow for many of these systems to complete their orbits. While azaleas flowered outside in the spring rain, astronomers inside presented results pointing to the startling conclusion that even the youngest stars are frequently surroundedby stellar companions. This realization was the product of painstaking observations by many different people using a host of clever techniques and new devices. That morning in Georgia, the separate works of these numerous researchers appeared magically to dovetail.
The finding that binary systems are at least as common for youngstars as for older ones might seem reasonable enough, but for astronomers it came as a shock. Most notions of double star formation had predicted that stellar companions are produced orcaptured well after a star has formed; hence, the youngest starswould be expected to exist singly in space. Such theories no longer bear weight. There remains, however, at least one idea forthe formation of double stars that holds up to the recent observations. It may be the sole explanation for why binary starsystems are so abundant in the universe.
The sun, a mature star, has no known stellar companions, even though most stars of its age are found in groups of two or more.In 1984 Richard A. Muller of Lawrence Berkeley Laboratory and his colleagues hypothesized that the sun is not truly a single starbut that it has a distant companion orbiting it with a period of about 30 million years. He reasoned that gravitational forces from this unseen neighbor could disturb material circling in the outermost reaches of the solar system, sending a shower of comets toward the inner planets every time the star neared. Muller suggested that this effect might explain periodic massextinctions: comets generated by the sun's companion would hit the earth every 30 million years or so and as with the demise ofthe dinosaurs would have wiped out much of life on earth. Because its approach would have sparked such widespread destruction, Muller called the unseen star "Nemesis."
Most scientists have not accepted Muller's interesting idea. Forone, the closest known stars (the Alpha Centauri triple star system, at a distance of 4.2 light-years) are much too far away to be bound to the sun by gravity. In fact, there is no astronomical evidence that the sun is anything other than a single star whose largest companion (Jupiter) is 1,000 times less massive than the sun itself. But living on a planet in orbit around a solitary sun gives us a distorted view of the cosmos; wetend to think that single stars are the norm and that double stars must be somewhat odd. For stars like the sun, this turnsout to be far from true.
In 1990 the late Antoine Duquennoy and Michel Mayor of the Geneva Observatory completed an exhaustive, decade-long survey of nearby binary stars. They considered every star in the sun's "G-dwarf" class within 72 light-years, a sample containing 164 primarystars that are thought to be representative of the disk of ourgalaxy. Duquennoy and Mayor found that only about one third of these systems could be considered true single stars; two thirdshad companions more massive than one hundredth the mass of the sun, or about 10 Jupiters.
Binary star systems have widely variable characteristics. Starsof some double G-dwarf systems may be nearly touching one another; others can be as far apart as a third of a light-year. Those in contact may circle each other in less than a day, whereas the most widely separated double stars may take tens ofmillions of years to complete a single orbit. Duquennoy and Mayorshowed that triple and quadruple G-dwarf stars are considerably rarer than double stars. They counted 62 distinct doubles, seven triples and two quadruple groupings. They further determined that each of the triple and quadruple sets had a hierarchical structure, composed of a relatively close double orbited by either a more distant single star (forming a triple system) or another close double star (forming a quadruple system). The separation between distant pairs needs to be at least five times the gap of the close doubles for the group to survive for long. Arrangements with smaller separations are named Trapezium systems, after a young quadruple system in the Orion nebula. These arrangements are orbitally unstable they will eventually flyapart. For instance, if the three stars of a triple system comeclose enough together, they will tend to eject the star of lowestmass, leaving behind a stable pair.
Double stars thus seem to be the rule rather than the exception. This conclusion does not, however, mean that planets must berare. A planet could travel around a double star system provided that it circles either near one of the two stars or far away fromboth of them. Imagine living on such a world orbiting at a safe distance from a tightly bound binary, where the two stars complete an orbit every few days. The daytime sky would contain apair of suns separated by a small distance. Sunrises and sunsetswould be fascinating to watch as first one and then the other glowing orb crossed the horizon. Other strange celestial configurations might also occur. If, for example, the planet orbited in the same plane as did two stars of equal mass, the two suns periodically would appear to merge as they eclipsed eachother, briefly halving the amount of combined sunlight reaching the planet.
The sun formed about 4.6 billion years ago and has about five billion years remaining of its so-called main-sequence lifetime. After it reaches the end of its main sequence, it will expand tobecome a red giant that will engulf the inner planets. This configuration will be somewhat akin to one that occurred early in the sun's history, when it extended far beyond its presentradius. At that time, before it had contracted to its currentsize, the sun was similar to the T Tauri class of stars that can be seen in those regions of our galaxy where stars are now forming. During its T Tauri stage, the sun's radius was about four times greater than its present measurement of some 700,000 kilometers. Still earlier, the protosun must have extended out to about 1.5 billion kilometers, or 10 times the distance betweenthe earth and the sun (that span, 150 million kilometers, is known as an astronomical unit, or AU).
Present-day T Tauri stars offer astronomers an opportunity to learn what the sun was like early in its evolution. The nearest T Tauri stars are in two locations, known as the the Taurus molecular cloud and the Rho Ophiuchus molecular cloud, both about 460 light-years from the earth. The fact that young stars are always embedded in such dusty concentrations of gas gives convincing testimony to their origin-stars are born from the contraction and collapse of the dense cores of molecular hydrogenclouds.
Because young stars are typically enshrouded by dust, astronomers usually have difficulty viewing them in visible light, no matterhow powerful the telescope. But these sites can be detected readily using infrared wavelengths that are characteristic of the emission from heated dust grains surrounding the nearby star. Progress in understanding the formation of stars has thus be endependent to a large extent on the development of detectors capable of sensing infrared radiation. At the 1992 meeting in Georgia, the first results were presented for several different infrared surveys specifically designed to detect companions tothe T Tauri stars in Taurus and Ophiuchus.
Andrea M. Ghez, now at the University of California at Los Angeles, and her colleagues Gerry F. Neugebauer and Keith Matthews, both at the California Institute of Technology, used a new indium antimony array camera on the five-meter Hale telescope to photograph the regions around known T Tauri stars at then ear-infrared wavelength of 2.2 microns. (Visible light has a wavelength between about 0.4 and 0.7 micron.) Using a so-called speckle imaging technique to minimize the noise introduced by fluctuations in the earth's atmosphere above the telescope, Ghezand her colleagues found that almost half of the 70 T Tauri starsin their sample showed stellar companions. For the limited range of separations considered, about 10 to 400 AU, this study indicated that for the youngest systems, binaries are twice as common as for main-sequence stars. Christoph Leinert of the Max Planck Institute for Astronomy in Heidelberg also presented results of a near-infrared speckle imaging survey. Leinert and his colleagues found that 43 of the 106 T Tauri stars they examined had nearby companions, again implying that binaries were much more common in these stars than in G-dwarf stars like oursun.
Hans Zinnecker and Wolfgang Brandner of the University of Wrzburg in Germany and Bo Reipurth of the European Southern Observatory in Chile used a high-resolution digital camera in combination with the European New Technology Telescope to image 160 T Tauri stars at an infrared wavelength (one micron). They uncovered 28 companions lying from 100 to 1,500 AU from the T Tauri stars, about a third more than circle around older, solar-type stars in that distance range.
Michel J. Simon of the State University of New York at Stony Brook, along with Wen Ping Chen (now at National Central University in Taiwan) and their colleagues, reported a novel way to find young double stars. When the moon passes over, oroccults, a distant star system, careful monitoring of the light received can reveal the presence of two or more sources, as first one and then another star slips behind the sharp edge of the lunar face. Simon and Chen's measurements detected companions much closer to T Tauri stars than was possible with infrared imaging. Their work again showed that a large fraction are binaries. Robert D. Mathieu of the University of Wisconsin employed a more traditional means for detecting close doublestars, the same as that used by Duquennoy and Mayor. Mathieu usedspectroscopic measurements of the periodic Doppler shift to show that some T Tauri stars have companions within 1 AU. Once more, closely spaced binaries proved more common in young T Tauri systems than for solar-type stars.
Search for a Theory
How did all these stellar companions come to be? Why did they form so abundantly and so early in their evolution? The wealth of observations of young stars presented in Georgia requires that binary stars must form well before even their pre-main-sequence (T Tauri) phase. Moreover, the finding that binaries are socommon demands that the mechanism generating them whatever it is must be very efficient.
In principle, a double star system could arise from two stars that pass close enough together so that one forces the other into a stable orbit. The celestial mechanics of such an event, however, requires the intervention of a third object to removethe excess energy of motion between the two stars and leave them trapped in a gravitationally bound system. But such three-bodyencounters are too rare to account for very many binary stars. Cathy J. Clarke and James E. Pringle of the University of Cambridge studied a more likely way that companion stars might have paired up. They investigated the gravitational coupling between two young stars that still had flattened disks of dust and gas surrounding them. That geometry would be far more common than three-body encounters and could, in theory, remove enough energy from the stars' motions. But in their analysis they found that such interactions are much more likely to end up ripping apartthe circumstellar disks than to result in one star's neatly orbiting with the other. So this embellishment seems to help little in explaining the existence of binary star systems.
Failure of the capture mechanism has forced most astronomers to think about processes that might form binary stars more directly. In fact, consideration of this notion goes back over a century. In 1883 Lord Kelvin proposed that double stars result from rotational fission. Based on studies of the stability of bodiesin rapid rotation, Kelvin suggested that as a star contracted, it would spin faster and faster until it broke up into a binary star. Astronomers now know that pre-main-sequence stars contractconsiderably as they approach the hydrogen-burning main sequence, but T Tauri stars do not rotate fast enough to become unstable. Furthermore, Kelvin's fissioning would act too late to explain the frequency of binaries among young stars. Richard H. Durisenof Indiana University and his colleagues showed that fission fails on theoretical grounds as well a reasonable calculation of this instability shows that the ejected matter would end up astrailing spiral arms of gas rather than as a separate cohesive star.
In contrast to the century-old fission theory, there is an idea for creating binary stars that is only a decade old, called fragmentation. This concept supposes that binary stars are born during a phase when dense molecular clouds collapse under their own gravity and become protostars. The obscuring gas and dustthen clear away, and a newly formed binary star (of the T Tauri class) emerges. In contrast to older theories of the birth of binary systems, fragmentation fully agrees with the latest observations of young stars.
The protostellar collapse that enables fragmentation occurs relatively suddenly in the scale of a several-billion-yearstellar lifetime; the event takes place in a few hundred thousandyears. This violent transformation of a diffuse cloud into acompact star thus offers a special opportunity for a single objectto break into several distinct members. Astrophysicists have identified two mechanisms that might operate. Very cold cloud scan fragment directly into binaries, whereas warmer clouds with substantial rotation can first settle into thin disks that later break up as they gain more mass or become progressively flattened.
A key objection to the fragmentation theory involved the distribution of matter in protostellar clouds. It was previously thought that this material was distributed according to a so-called power law. That is, there would be an extremely high concentration of material near the center of the cloud and arapid decrease in density with distance. This objection appears, however, to have been removed recently by high-resolution radio observations made using submillimeter wavelengths. Last year Derek Ward-Thompson of the Royal Observatory in Edinburgh and hiscolleagues determined the distribution of material inside several precollapse clouds. They found that the density follows a Gaussian (bell-shaped) distribution rather than a power law. Hence, matter would be less tightly concentrated toward a centralpoint when the star system began to form. Elizabeth A. Myhill, then at the University of California at Los Angeles, and I had shown separately that the high density at the center of a cloudthat follows a power law makes it almost impossible for a secondor third star to coalesce. It proves much easier for fragmentation to occur with an initial Gaussian distribution.
Astrophysicists can predict whether multiple fragments will ultimately form by solving the set of equations that govern the flow of gas, dust and radiation in a protostellar cloud. The calculations are sufficiently complex to require accurate software and a powerful computer for their solution. I began modeling the collapse of dense clouds with Gaussian density profiles in 1986 and found that fragmentation could readily occur provided certain conditions were met. As long as a Gaussian cloudhas sufficient rotation to give the binary system the angular momentum it requires and the precollapse material is cold enough (less than 10 kelvins) to make its thermal energy less than about half its gravitational energy, the cloud will fragment during its gravitational contraction. The conditions appear to be nothingout of the ordinary for the clouds found in stellar nurseries.
Whether a binary, triple or quadruple system eventually forms depends on many details, including the three-dimensional shape ofthe original cloud, how lumpy it is and the precise amount of thermal and rotational energy available. In general, prolate, or football-shaped, clouds tend to form bars that fragment into binary systems, whereas more oblate, or pancake-shaped, clouds flatten to disks that later fragment into several members.
The collapse is thought to occur in two separate steps. The first phase generates protostars with a radius on the order of 10 AU. Thus, the first phase of fragmentation can produce only binary systems with separations of about 10 AU or larger. These bodies then undergo a second collapse to form the final protostars of stellar dimensions. Ian A. Bonnell and Matthew R. Bate of the University of Cambridge have shown that fragmentation can happen during the second collapse phase as well, and this process can lead to the formation of protostellar cores separated by distances comparable with those of the closest main-sequencestars. Fragmentation appears to be capable of generating the entire range of separations observed in young binary stars, from the closest to the widest systems
Brown Dwarfs and Giant Planets
What about finding companions of even lower mass? Duquennoy and Mayor produced evidence that as many as 10 percent of solar-typestars are bound to brown dwarfs that is, they have stellar companions with masses from 0.01 to 0.08 times the mass of the sun. Brown dwarfs are too small to ignite hydrogen the way the sun does but could be massive enough to burn deuterium soon after formation. After that, their radiation would cease, and they would become cool and extremely difficult to detect. Although the evidence offered by Duquennoy and Mayor is intriguing, there is as yet no confirmed example of a brown dwarf star, in spite of the many efforts to detect one. The search is also on for planetary companions, although again astronomers have yet found no convincing candidates. But in the next decade, experimental techniques should improve to the point that planets the size of Jupiter could be detected (or else demonstrated not to exist) around a number of nearby stars. Whether it is reasonable to examine binaries or restrict the search to single stars like thesun is an open question; astronomers will probably target some of both in their ongoing effort to uncover a planetary system comfortingly similar to our own.