1

Detection of Substellar-Mass Objects

How unique is our own planetary system? What is the likelihood of life elsewhere in the cosmos? These questions have always been of great interest to scientists and nonscientists alike. The recent discovery of 51 Pegasi B1.2 and a dozen or more additional objects as massive or more so than the planet Jupiter in orbit around nearby stars has greatly intensified interest in these questions. These discoveries, combined with the identification of the first bona fide brown dwarf, an object intermediate in mass between a star and a planet, have brought the study of substellar-mass objects (SMOs) into a new phase of observational and theoretical investigation.

Although the intellectual linkage between study of SMOs and the question of the frequency of planetary systems is a firm one, for various programmatic reasons the future investigation of SMOs has not been carefully thought through to date. Past reports by the National Research Council's (NRC's) Committee on Planetary and Lunar Exploration (COMPLEX) 3 and NASA advisory groups4 on detection and study of extrasolar planets have concentrated primarily on the study of SMOs as the first step toward detecting Earthlike planets around other stars, the principal goal of NASA's new “Origins” initiative. Comparatively little attention has been given to what these objects could tell us about high-priority questions in astronomy and the planetary sciences.

The demonstrated ability to observe and study SMOs is significant for reasons additional to and unconnected with extrasolar planets. The abundance of low-luminosity, low-mass objects—sometimes called massive compact halo objects (MACHOs)—places constraints on models for the nature of the dark matter on a number of astrophysical scales, both directly (in terms of the SMO contribution) and indirectly through their constraints on star formation models. Attempts have been underway for several years to determine the galactic and extragalactic mass contribution of SMOs by detecting the microlensing effect (i.e., the gravitational lensing caused by an individual object) of such objects on background stars.5

The study of brown dwarfs, such as the prototypical example, Gliese 229B, pushes the technological challenges that must be met for the detection of planetary systems around other stars. Although some may regard the distinction between a planet and a brown dwarf as purely a question of semantics, the workshop's steering group adopted a formal definition based on the physics of their formation. A brown dwarf is an object formed in the same manner as a star by a

1  

M. Mayor and D. Queloz, “A Jupiter-Mass Companion to a Solar-Type Star,” Nature, 378: 355, 1995.

2  

A.P. Hatzes, W.D. Cochran, and E.J. Bakker, “Further Evidence for the Planet Around 51 Pegasi,” Nature, 391: 154, 1998.

3  

National Research Council, Space Studies Board, Strategy for the Detection and Study of Other Planetary Systems and Extrasolar Planetary Materials: 1990-2000, National Academy Press, Washington, D.C., 1990.

4  

C.A. Beichman (ed.), A Roadmap for the Exploration of Neighboring Planetary Systems, JPL 96-22, Jet Propulsion Laboratory, Pasadena, Calif., 1996.

5  

B. Paczynski, “Gravitational Microlensing in the Local Group,” Annual Reviews of Astronomy and Astrophysics, 34: 419, 1996.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects 1 Detection of Substellar-Mass Objects How unique is our own planetary system? What is the likelihood of life elsewhere in the cosmos? These questions have always been of great interest to scientists and nonscientists alike. The recent discovery of 51 Pegasi B1.2 and a dozen or more additional objects as massive or more so than the planet Jupiter in orbit around nearby stars has greatly intensified interest in these questions. These discoveries, combined with the identification of the first bona fide brown dwarf, an object intermediate in mass between a star and a planet, have brought the study of substellar-mass objects (SMOs) into a new phase of observational and theoretical investigation. Although the intellectual linkage between study of SMOs and the question of the frequency of planetary systems is a firm one, for various programmatic reasons the future investigation of SMOs has not been carefully thought through to date. Past reports by the National Research Council's (NRC's) Committee on Planetary and Lunar Exploration (COMPLEX) 3 and NASA advisory groups4 on detection and study of extrasolar planets have concentrated primarily on the study of SMOs as the first step toward detecting Earthlike planets around other stars, the principal goal of NASA's new “Origins” initiative. Comparatively little attention has been given to what these objects could tell us about high-priority questions in astronomy and the planetary sciences. The demonstrated ability to observe and study SMOs is significant for reasons additional to and unconnected with extrasolar planets. The abundance of low-luminosity, low-mass objects—sometimes called massive compact halo objects (MACHOs)—places constraints on models for the nature of the dark matter on a number of astrophysical scales, both directly (in terms of the SMO contribution) and indirectly through their constraints on star formation models. Attempts have been underway for several years to determine the galactic and extragalactic mass contribution of SMOs by detecting the microlensing effect (i.e., the gravitational lensing caused by an individual object) of such objects on background stars.5 The study of brown dwarfs, such as the prototypical example, Gliese 229B, pushes the technological challenges that must be met for the detection of planetary systems around other stars. Although some may regard the distinction between a planet and a brown dwarf as purely a question of semantics, the workshop's steering group adopted a formal definition based on the physics of their formation. A brown dwarf is an object formed in the same manner as a star by a 1   M. Mayor and D. Queloz, “A Jupiter-Mass Companion to a Solar-Type Star,” Nature, 378: 355, 1995. 2   A.P. Hatzes, W.D. Cochran, and E.J. Bakker, “Further Evidence for the Planet Around 51 Pegasi,” Nature, 391: 154, 1998. 3   National Research Council, Space Studies Board, Strategy for the Detection and Study of Other Planetary Systems and Extrasolar Planetary Materials: 1990-2000, National Academy Press, Washington, D.C., 1990. 4   C.A. Beichman (ed.), A Roadmap for the Exploration of Neighboring Planetary Systems, JPL 96-22, Jet Propulsion Laboratory, Pasadena, Calif., 1996. 5   B. Paczynski, “Gravitational Microlensing in the Local Group,” Annual Reviews of Astronomy and Astrophysics, 34: 419, 1996.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects process of gravitational collapse and fragmentation. A planet, however, is built up by accretion in a circumstellar disk of gas and dust. Although it is generally assumed that the average planet will be of lower mass than the average brown dwarf, the upper mass limit for planet formation and the lower mass limit for brown dwarf formation are not yet known. Indeed, these mass limits may overlap, making it very difficult to know whether a particular object is a planet or a brown dwarf. An alternative distinction is sometimes drawn between planets and brown dwarfs: the former is defined as having a mass less than 13 Jupiter masses (MJ) and is, hence, unable to undergo deuterium fusion. Coincidentally, this is in the general mass regime of ~10 to 20 MJ thought to be at the upper end of the planet formation spectrum and at the bottom of the range for starlike formation by direct fragmentation. STATUS OF CURRENT INDIRECT SEARCHES The Radial-Velocity Technique The indirect detection of substellar companion objects to stars has produced impressive results during the 1990s, with the discovery of about 20 objects below the stellar-mass limit. Most of these new discoveries are a direct consequence of major improvements in the precision with which it is possible to measure slight variations in a star's radial velocity (i.e., the speed with which the star approaches or recedes from us). In the mid- 1980s, the best routine velocity precision was about 500 m/s; now the state of the art is 3 m/s.6 The stellar radial velocity will be observed to vary as the star and its brown dwarf (or planetary) companion each execute symmetric reflex orbits around the center of mass (or balance point) of the system. Thus, from the detection of the stellar orbital motion, researchers can infer the presence of the unseen companion object and compute a lower limit on its mass. There are now about a half dozen research groups in the world that are capable of routinely measuring radial velocities with precisions in the range of 3 to 15 m/s. Because the Sun's reflex orbital velocity caused by Jupiter is about 12 m/s, high precision is needed to detect Jupiter-mass objects. The objects detected around other stars so far range from about 0.5 MJ7 all the way up to the stellar-mass limit of about 80 MJ (0.08 solar masses). Researchers do not yet have enough information to know which of these objects are planets or brown dwarfs. Some argue that the distribution of orbital eccentricities as a function of orbital period implies that virtually all of the substellar companions are brown dwarfs (some, by virtue of the sin i ambiguity, are in fact stars, as noted below). Others contend that this distribution is the result of early dynamical evolution of the system, rather than the mechanism of formation. Data reported at International Astronomical Union Colloquium 170 in June 1998 may shed some light on this issue. Nearly a decade of high-precision radial-velocity observations of some 200 stars by several groups have revealed seven companion objects with M sin i less than 5 MJ, one object with M sin i between 5 and 10 Jupiter masses, and no objects with M sin i between 10 and 80 MJ. These results imply that 10- to 80-MJ SMOs are found much less frequently within 5 AU of their parent star than are SMOs with mass less than 5 MJ. A lower 6   G.W. Marcy and R.B. Butler, “Detection of Extrasolar Giant Planets,” Annual Reviews of Astronomy and Astrophysics, 36: 57, 1998. 7   This is the lower limit to the mass due to an ambiguity in the inclination of the planet's orbit relative to our line of sight.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects precision survey of about 500 stars by the Swiss group found about 10 companions with M sin i in the range of 10 to 80 Mj,8 and a similar size survey by researchers at the Harvard-Smithsonian Center for Astrophysics found one object between 5 and 10 MJ.9 Data from the Hipparcos satellite, however, show that at least seven of the companions found in the Swiss survey have orbital inclinations such that sin i is small, rendering them hydrogen-burning, stellar companions rather than SMOs. Thus, SMOs from 10 to 80 MJ occur with a frequency of less than 1% within 5 AU, which is much less than that of SMOs with masses less than 10 MJ. The key to resolution of this question is to attempt to detect a second low-mass companion object around these stars. Planets should form in regular systems with several planets, whereas stars should form in hierarchical multiple systems. Several new, large-scale radial-velocity surveys have been started recently that will greatly expand the sample of stars surveyed for low-mass companion objects. These new surveys will help to provide the data necessary to resolve the formation mechanism for these bodies. Astrometric, Photometric, and Microlensing Techniques There are several alternative techniques for the indirect detection of SMOs. These include astrometry, which measures the change in position of a star in the sky due to its reflex orbit around the center of mass of the system, and photometry, which measures the dimming of the starlight when the companion object passes in its orbit between Earth and the star (for those systems with the orbital plane nearly along our line of sight to the star).10 A third, highly promising, technique is microlensing.11 This technique takes advantage of the transient, gravitational lensing that will occur when an object (the “lens”) in the halo or bulge of the Milky Way passes between Earth and a more distant star (the “source ”). The presence of a companion, in orbit, about the lensing star will create a characteristic perturbation in the light curve of the microlensing event. Neither astrometry nor photometry has yet crossed the threshold of being a sufficiently mature technique to be able either to produce detections of substellar companions or to place interesting limits on non-detections in current large-scale surveys. Microlensing is a well-proven technique for finding binary systems and, as such, its extension to the detection of brown dwarfs and extrasolar planets should be relatively straightforward. More details about microlensing and the other applications to the study of SMOs can be found in Chapter 5. The precision of classical, single-aperture astrometry has improved significantly in recent years and now is beginning to approach 0.0001 arc sec (0.1 milliarc sec) on large (5- to 10-m-class) telescopes. If this precision can be maintained over multiyear intervals, astrometry holds the promise over the next decade of detecting Jupiter-mass objects in orbit at 4 to 5 AU from the nearest 30 solar-type stars. 8   M. Mayor, D. Queloz, and S. Udry, “Mass Function and Orbital Distributions of Substellar Companions, ” in Brown Dwarfs and Extrasolar Planets, R. Rebolo, E.L. Martin, and M.R. Zapatero Osorio (eds.), ASP Conference Series 134, Astronomical Society of the Pacific, San Francisco, Calif., 1998, page 140. 9   D.W. Latham et al., “The Unseen Companion of HD114767—A Probable Brown Dwarf,” Nature, 339: 38, 1989. 10   W.J. Borucki and A.L. Summers, “The Photometric Method of Detecting Other Planetary Systems,” Icarus, 58: 121, 1984. 11   S. Mao and B. Paczynski, “Gravitational Microlensing of Double Stars and Planetary Systems. ” Astrophysical Journal, 374: L37, 1991.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects Significantly higher astrometric precision can be achieved through interferometry, in which the light from multiple separated apertures is combined coherently. With apertures arranged on a 100-meter baseline, precision of tens of microarc sec are achievable. The Palomar Testbed Interferometer (PTI) has already demonstrated the technology to do this, and NASA is now developing an interferometer for the Keck Telescopes (see presentation by C. Beichman in this chapter). Several other optical or near-infrared interferometers either are newly commissioned or are under construction. These facilities include the Navy Prototype Optical Interferometer (NPOI), the Center for High Angular Resolution Astronomy (CHARA), and the European Very Large Telescope Array. Photometric observation of planetary transits is the only indirect detection technique that offers the ability to measure the radius of the companion object; it can also measure the size of the orbit, but not the mass of the object. When this technique is used from ground-based telescopes, the precision of the measurements is limited by variations in atmospheric extinction and scintillation. The method should easily be able to detect Jupiter- and brown dwarf-size objects in relatively short-period orbits. Indeed, several groups are now exploring this possibility. If applied to a space-based telescope, this method should be able to achieve the photometric precision necessary to detect Earth-size objects around a very large sample of stars. These techniques exhibit an exciting complementarity. Radial-velocity searches explore the space near the parent star, whereas astrometry is more sensitive to companion objects in wider orbits (see Figure 1.1). Astrometry is best applied to the nearest stars, most of which are too faint for radial-velocity techniques. Photometric detection measures both the physical size of the companion and its orbital inclination. When combined with radial-velocity data on the same system, photometric detection allows computation of the mass and the density of the object, allowing a direct test of theoretical models. Microlensing is also highly complementary to other approaches. While radial-velocity studies are most sensitive to the detection of close companions and astrometric searches are most sensitive to the detection of widely separated companions, microlensing is most sensitive to those with intermediate (i.e., 2- to 6-AU) spacings. Microlensing also stands out from the other techniques in that it can detect objects regardless of luminosity. DIRECT DETECTION OF SMOs In addition to the indirect methods discussed above, the intrinsic luminosity of SMOs, particularly brown dwarfs, may be directly detected using a number of different techniques. Much current research is focused on using ground-based telescopes to detect the long-wavelength emissions from brown dwarfs. In the not too distant future, these efforts will be supplemented by observations using new ground-based facilities (e.g., the Keck Interferometer) and space-based systems, such as the Wide-Field Infrared Explorer (WIRE), the Space Infrared Telescope Facility (SIRTF), and the Space Interferometer Mission (SIM). In the more distant future, it may be possible to detect Jupiter-size objects (see presentation by R. Angel in this chapter). A Next Generation Space Telescope (NGST), equipped with a cooled, active primary mirror 6 meters or more across, would be able to detect extrasolar giant planets through their anomalously high thermal emission at 5 μm. A Jupiter-size planet could be resolved from its parent star in a solar-system twin at a distance of 10 parsecs. In the much more distant future, space-based interferometers may enable imaging and spectroscopic study of Earthlike planets in orbit about nearby stars. 12

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects FIGURE 1.1 Brown dwarfs and giant extrasolar planets discovered to date (together with Jupiter and Saturn) are plotted as functions of their masses and their separation from their parent star. The diagonal dashed and dashed-dotted lines show the approximate sensitivity of current radial-velocity and astrometric-search techniques, respectively. Objects detected by the radial-velocity method generally have only lower limits on their masses because the inclination of their orbit to the line of sight is unknown. Objects to the right of the vertical double lines are not bound to any star, and the objects to the left of the vertical single line have been circularized because of tidal effects. Objects less massive than some 13 and 80 MJ are incapable of sustaining deuterium and hydrogen fusion reactions, respectively. Illustration courtesy of Alan Boss, Carnegie Institution of Washington.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects In the near future, however, attention should be restricted to the direct detection of brown dwarfs. Strategies for achieving this fall into two categories: Direct detection of isolated brown dwarfs; and Direct detection of companion brown dwarfs. The first strategy is technically the simplest to implement. The main challenge is knowing where in the sky to look. Using the second strategy, knowing where to look is obvious—at nearby stars—but the technical challenge is to resolve the faint brown dwarf against the bright glare of its companion star. Isolated Brown Dwarfs The 2-Micron All Sky Survey (2MASS) and the Deep Infrared Survey of the Southern Sky (DENIS) have already identified more than a dozen brown dwarf candidates (see presentation by J. Stauffer in this chapter). These tentative identifications, based on the color characteristics of objects with a brown dwarf's low effective temperature, have subsequently been confirmed by the detection of lithium absorption lines in their spectra. Lithium is too readily destroyed in thermonuclear reactions to be abundant in stars, and, thus, its presence is an indication that an object is substellar (see Chapter 2). Extrapolations based on current results from 2MASS and DENIS suggest that they will eventually find a few hundred to a few thousand brown dwarfs. The brown dwarfs found by infrared surveys such as DENIS and 2MASS will not necessarily be typical of the total population of such objects. Infrared surveys preferentially find warm brown dwarfs. Only a handful of objects as cool as Gliese 229B will be identified in these surveys. Discovery of a significant population of cool objects will have to await the launch of the WIRE and SIRTF missions in 1998 and 2001, respectively. Companion Brown Dwarfs Gliese 229B, the prototypical brown dwarf in a binary system, was discovered during a systematic campaign to obtain high-dynamic-range imaging of nearby stars (see Figure 1.2).13 Development of this imaging technique—the ability to detect very faint objects adjacent to very bright ones—is fundamental to the search for extrasolar planets. The greatest advances in this area are expected to come primarily from the use of adaptive-optics systems on large ground-based telescopes. The primary advantage of these systems is their ability to reduce the size of the core of the bright star's image. Current adaptive-optics systems do not, however, 12   For a recent review of relevant issues, see, for example, N. Woolf and J.R. Angel, “Astronomical Searches for Earth-Like Planets and Signs of Life,” Annual Reviews of Astronomy and Astrophysics, 36: 507, 1998. 13   T. Nakajima, B.R. Oppenheimer, S.R. Kulkarni, D.A. Golimowski, K. Matthews, and S.J. Durrance, “Discovery of a Cool Brown Dwarf,” Nature, 378: 463. 1995.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects FIGURE 1.2 Since the first direct observation of Gliese 229B in 1995, this brown dwarf has been imaged by a variety of techniques. Some of the best images resolving Gliese 229B (the bright spot in the five o'clock position) from its bright companion, Gliese 229A (center), were obtained at NASA's Infrared Telescope Facility using the NSFCam and the University of Arizona's Co-Co cryogenically cooled infrared coronagraph. The top frame shows the raw data. The second frame is an average of several images processed using the technique of methane difference imaging to remove the residual stellar halo seen in the first frame. This technique involves taking images in wavelengths corresponding to a methane absorption band and the continuum. Subtracting these images from each other highlights the cool, methane-bearing Gliese 229B while suppressing Gliese 229A, in which methane is absent. The bottom frame is similarly processed except that the residual halo was removed using a technique devised by Christ Ftaclas. The latter technique involves subtraction of synthetic stellar profiles constructed by azimuthal averaging of each quadrant. Each quadrant must be treated separately because mispointing of the telescope can cause asymmetries in a coronagraphic image. Both techniques have their strengths and weaknesses. The methane differencing does a much better job of removing the diffraction spikes (caused by the supports for the telescope's secondary mirror), whereas quadrant averaging produces a flatter image in the regions between the spikes. Images courtesy of Robert H. Brown and David Trilling, University of Arizona, and Christ Ftaclas, Michigan Technical University.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects significantly decrease the extended faint halo of scattered light surrounding this core. Thus, adaptive optics will only reveal faint close companions that are not lost in their bright companion's halo of scattered light. Many such systems are either operational (e.g., on the Canada-France-Hawaii Telescope and the European Southern Observatory) or are about to become operational at several observatories in the United States (e.g., Keck, Palomar, and the MMT) and in Europe. A wave of discoveries of objects as cool as or cooler than Gliese 229B is likely to follow the perfection of this technology. The Hubble Space Telescope (HST) is playing an important role in the search for brown dwarfs. HST's Near-Infrared Camera and Multi-Object Spectrometer (NICMOS) is currently being used to conduct imaging surveys to detect companion brown dwarfs. In one such effort, all stars within 10 parsecs are being examined. In other surveys, many nearby and young stars are being imaged either directly or with NICMOS 's coronagraph. These observing programs have sufficient sensitivity to find objects as cool as 600 K and separated by distances in excess of 1 arc sec from their companion stars. Observers are confident that these efforts will unearth cool brown dwarfs—objects of considerable interest to the planetary scientists given the resemblance between the spectrum of Jupiter and that of the cool brown dwarf Gliese 229B. Interferometry will play an increasing role in studies of both companion brown dwarfs and hot Jupiters (i.e., objects such as 51 Pegasi B, identified as a result of radial-velocity surveys). These objects are relatively bright but lie so close to their companion stars that adaptive optics is required to resolve them. Fortunately, the long baseline of the Palomar Testbed Interferometer and the Keck Interferometer will enable researchers to image these objects. In addition, the development of spectral-line interferometric techniques—now well under way—will enable these facilities to obtain crude spectra. In the longer term, application of indirect techniques such as radial-velocity studies (whose time baseline is continuously growing) and astrometric surveys of nearby stars (using the Keck Interferometer and the Palomar Testbed Interferometer) will greatly increase the number of brown dwarf candidates available for study. Indeed, the new interferometers may allow sampling of several hundred nearby stars to a level sufficient to detect a Uranus-mass object 5 AU from its parent star (see presentation by R.W. Noyes in this chapter). In the somewhat longer term, these ground-based studies will be supplemented by those of the Space Interferometer Mission (SIM). This mission will carry out a large survey of nearby stars, both directly as targets and indirectly as reference objects, with a sensitivity approaching that required for direct detection of Earth-mass objects. Thus, SIM can very easily identify brown dwarfs and hot Jupiters. However, spectroscopic studies of these objects will require interferometers with longer baselines.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects SEARCH STRATEGIES FOR ISOLATED BROWN DWARFS USING CURRENT GROUND-BASED FACILITIES John Stauffer Harvard-Smithsonian Center for Astrophysics Perhaps slightly less well publicized than the recent detection of substellar-mass companions to nearby stars (such as the companions to 51 Pegasi and Gliese 229), and lagging behind by more or less 1 year in time, the field of detection of “free-floating” substellar objects has also recently been converted from a field of upper limits to one of exciting discoveries. The first journal paper announcing the confirmation of a free-floating brown dwarf via detection of lithium in its spectrum was published in October 1996 (the object was a member of the Pleiades designated as PPL 15). The first paper announcing lithium detection in a free-floating brown dwarf that was not a member of an open cluster was published in November 1997. However, this field is evolving rapidly also, and there are now at least 10 confirmed brown dwarfs in open clusters and more than 5 confirmed, free-floating, field brown dwarfs. Field brown dwarfs have been discovered via: Proper-motion surveys with photographic plates, searching for rapidly moving, red objects; Deep optical multifilter imaging; and Deep optical-infrared imaging using one optical band and at least two infrared bands. The first two methods can succeed, but they require a large follow-up effort because purely optical colors do not discriminate well between low-mass stars and brown dwarfs. The third method, as employed by the Deep Infrared Survey of the Southern Sky (DENIS) and the 2-Micron All-Sky Survey (2MASS), is very efficient and has a good success rate, with six confirmed brown dwarfs after surveying less than 1% of the sky. Because brown dwarfs are brighter and warmer when they are young and because these surveys are flux limited, they are biased toward identifying relatively young, relatively high-mass brown dwarfs. As a result, none of the candidates that have been confirmed to date appear to be as cool as Gliese 229B, and all probably have masses greater than about 0.04 solar masses. It was recognized 10 years ago that young, nearby, open clusters would be good sites to search for brown dwarfs. The youth of these clusters allows these surveys to be done using multifilter, optical imaging, making it possible to survey a significant fraction of the cluster to quite deep limits. Due to a combination of a favorable age, distance, and position in the sky, by far the best-studied cluster has been the Pleiades. With perhaps one exception, all of the confirmed open-cluster brown dwarfs are Pleiades members. The existing photometry and spectroscopy now define the “lithium depletion” point in the Pleiades to within ±0.1 magnitude, allowing a more precise determination of the age of the cluster than by any other means. The mass function of the Pleiades is rising slowly down to the limits of the current surveys (~0.04 solar masses). The future of both the field and cluster searches for substellar objects is bright. Upon completion of the 2MASS and DENIS projects (around 2001), it is likely that several hundred field brown dwarfs will have been identified. By 2001, I expect that brown dwarfs will have been confirmed in several open clusters and the “lithium age ” for the cluster determined. This should not only provide empirical calibration for the theoretical models of low-mass stars, but by

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects giving accurate cluster ages should also provide an important calibration of high-mass stellar models. Hundreds of free-floating brown dwarfs will be identified in the next few years—the methodology and the instrumentation to do so are now in place. The problems ahead will be to ensure that the manpower to analyze and follow up these identifications is available, to continue to provide access to big telescopes (such as Keck) to obtain the necessary spectroscopy, and to develop appropriate theoretical models to allow an understanding of how biases inherent in these surveys shape the masses and ages of the detected objects. SEARCH AND ANALYSIS STRATEGIES FOR SUBSTELLAR AND PLANETARY COMPANIONS BY DIRECT DETECTION Roger Angel Steward Observatory, University of Arizona The observational techniques needed to search for substellar-mass objects by direct detection are basically the same as those needed to find extrasolar planets. For both types, direct detection will have its greatest value when used for spectroscopic analysis. When low-resolution infrared spectra of terrestrial planets around other stars can be realized, the potential is extraordinary. Habitability could be established on the basis of temperature and atmospheric water. Abundant, primitive life based on organized molecular structure might reveal itself, as on Earth, by an atmospheric composition modified in ways unlikely by inorganic processes. Such spectral analysis is the goal of the Terrestrial Planet Finder (TPF) mission. The techniques to isolate the emission from companions both very close to and enormously fainter than their star need to be developed and tested. We can anticipate this happening in an evolutionary process, from the current use of the Hubble Space Telescope and single large telescopes on the ground, through ground-based interferometers to the Next Generation Space Telescope (NGST) telescopes and ultimately the TPF. As improvements are made, companions of progressively smaller mass and closer separation will become accessible. Thus at each stage we can expect new scientific discoveries. Ground-based telescopes will likely remain the most important observational tools for a number of years. The new methods of adaptive-optics correction will work especially well when applied to fields close to bright stars and are thus very powerful for direct detection of companions. By correcting their atmospheric wavefront errors to near-Hubble levels of accuracy, single telescopes such as the 6.5 m in the MMT and the Magellan twins can be used to search the whole sky for Jupiter-size companions in reflected light, to distances of ~8 parsecs. Sensitivity to the thermal emission from companion objects will be limited by the telescope emission, but will still be sufficient to allow direct detection with 8-m-class telescopes of 5-Gy companions down to the brown-dwarf/superplanet boundary. Optical interferometry from the ground will be able to detect companions too close to be resolved with single apertures, and to detect zodiacal dust in extrasolar planetary systems. The latter measurements are crucial, for if the dust underlying planets is found to be much thicker than that in the solar system, direct detection with TPF will be compromised. The Keck interferometer will resolve the closest, star-heated companions (0.1 AU at 10 parsecs), while the Large Binocular Telescope (LBT) in nulling mode will obtain the best measure of the dust emission at #DXGT#0.5 AU, the component underlying terrestrial planets.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects In space, an NGST with D ≥ 6 meters and with a cooled, active primary built to the precision accuracy already demonstrated in an ultra-lightweight prototype will be able to detect extrasolar giant planets through their anomalously high thermal emission at 5 μm. A Jupiter-size planet could be resolved from a star in a solar-system twin at 10 parsecs. Sensitivity to planets below 1 μm could be obtained by incorporating additional active-optics correction in the focal-plane instrumentation. For bright nearby stars, this could be expected to yield residual wavefront errors as small as 2 nm rms. With advanced coronagraphic optics, the optical halo would be reduced enough to image terrestrial, as well as giant, planets in systems to 8 parsecs distance. If the very closest stars, τ Ceti and three others, have terrestrial planets, even spectroscopy would be possible at resolution λ/Δλ = 30, with the sensitivity to detect CO2, water, and oxygen features near 1 μm. The reflected light of giant extrasolar planets could be detected to 10 parsecs and analyzed spectroscopically for the strong features of methane and ice seen in Jupiter and Saturn. A robust search for terrestrial planets of ~100 nearby solar stars must await the TPF. It will take full advantage of the improved contrast and spectroscopic richness of the thermal infrared. Direct detection of thermal emission to 15 parsecs will be made possible by use of a ~75-m-long interferometer that both cancels starlight by destructive interference and allows for aperture synthesis to map planet positions. Such image reconstruction requires spectral separation, and so the same interferometer will be able to obtain simultaneously spectra of all the planets around a star, given integrations of up to several months. The strong greenhouse gas features of CO2, H2O, and O2 as seen in Earth's spectrum would be detectable. INDIRECT DETECTION TECHNIQUES PRESENT AND FUTURE Robert W. Noyes Harvard-Smithsonian Center for Astrophysics We review three different classes of detection technique, indicating their current capabilities and future potential: Radial velocity measurements are being pursued by about a dozen research groups, with typical precision of 10 m/s or better. The best performance achieved to date is about 3 m/s. This work has yielded about 20 detections of substellar-mass companions in the past few years, with minimum masses ranging from 0.5 to about 60 Jupiter masses. The surprising discovery of a number of “hot Jupiters” in close orbits about Sunlike stars means that radial velocity (RV) signals on the order of 50 m/s are not rare. Such variations are readily detectable with stable spectrometers at moderate-size (2-m-class) telescopes, and surveys are now under way in both hemispheres designed to survey as many as 1000 F, G, and K main sequence stars for such variations. This is important, because the current tally of some 20 detections is still insufficient for reasonable inferences on the origin or evolutionary history of these objects. At the same time, it is important to use larger telescopes to detect smaller-amplitude variations produced by long-period and/or lower-mass companions, and by second companions in systems with known giant-planet companions. Finally, it is important to extend the survey to yield information on the frequency of companions to stars of different masses, ranging from A-type stars to late-M stars. This poses the requirement to carry out PRV surveys with a mix of moderate-aperture (2-m-class) and large (6- to 10-m-class) telescopes; the former can make very frequent observations to determine

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects orbital parameters of brighter (F, G, and early-K), short-period systems, while the latter can make less frequent observations of stars that are fainter (late-K and M) and/or have lower-amplitude orbits due to longer-period and/or lower-mass companions. These complementary programs should be carried out for at least a decade or two. There is no need to go to space for this work. Astrometric detection is complementary to RV detection. Ground-based astrometry using single-aperture systems can detect orbital wobbles in the 0.1- to 1-milliarc sec (mas) range. and in the coming few years may well achieve firm astrometric detections of a few Jupiter-mass objects. New interferometers, such as the Palomar Testbed Interferometer (PTI) and the Keck Interferometers, should be able to push the limits of detection down to perhaps 30 microarc sec. This should allow sampling of some 200 nearby stars to a level sufficient to detect a Uranus-mass object at 5 AU, and will probably make single-aperture astrometry obsolete. To detect Earthlike planets by astrometry, however, probably requires going to space. The Space Interferometry Mission (SIM) should achieve a wide-angle precision of about 4 microarc sec, and narrow-angle precision as good as 1 microarc sec; the latter figure would allow detection of a 1-Earth-mass planet in a 1-AU orbit about a 0.3-solar-mass star at a distance of 3 parsec. Photometric detection of transits by giant planets should readily be achievable from the ground (1% light modulation), and several groups are attempting this very low cost approach. To detect Earthlike planets, however, it is necessary to go to space. The proposed Kepler mission might detect about 500 Earth-sized planets, if the solar system is typical. Non-imaging detection of emission from a hot Jupiter is possible using slight variants of all three of the techniques above. High-resolution spectroscopy with stable spectrometers may be able to detect the Doppler-shifted reflected light from close-in giant companions like 51 Pegasi B making use of known orbital parameters of the system. Astrometric studies using phase-difference interferometry can detect wavelength-dependent photocenter shifts due to a low-mass companion with different spectral signature from the primary. Photometric detection of orbital-phase-variations of the reflected-light signal from close-in giant planets is feasible with the Kepler mission. The evolution of approaches from ground-based photometry, radial velocity, and astrometry to space platforms such as SIM and Kepler make a natural and logical progression in capability, scientific return, and cost. Adequate support of the earlier stages of this progression is necessary to pave the way for later stages. DETECTION OF SUBSTELLAR-MASS OBJECTS FROM SPACE: FROM BROWN DWARFS TO TERRESTRIAL PLANETS Charles Beichman Infrared Processing and Analysis Center, Jet Propulsion Laboratory The long-term goal of NASA's Origins program is the detection and characterization of terrestrial planets. While the detection of the first Earthlike planet beyond our solar system lies many years and many daunting technical challenges in the future, the task of finding objects smaller than stars, orbiting other stars as planets or floating freely in space as brown dwarfs, is already well under way using a combination of ground- and space-based observatories.

OCR for page 6
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects By definition, neither brown dwarfs nor planets attain a mass high enough to initiate sustained thermonuclear burning. The manner of their birth serves to define a difference between the two types of substellar mass objects. A brown dwarf forms as a collapsing fragment of a molecular cloud exactly as do more massive stars. Planets, on the other hand, form via an agglomeration of gas, ice, or more refractory materials in a protostellar disk. Astronomers are finding both planets and brown dwarfs at an accelerating pace. Searches in clusters like the Pleiades, as well as the Deep Near-Infrared Survey of the Southern Sky (DENIS) and 2-Micron All Sky Survey (2MASS), are now identifying large numbers of brown dwarfs, as confirmed by the presence of easily-destroyed lithium in follow-up optical spectra. Since near-infrared searches can detect only the hottest brown dwarfs (2500#DXGT#T#DXGT#1000 K), these brown dwarfs are the most massive or the youngest of the class of substellar objects. Smaller or older objects will be detectable only in long-wavelength surveys like that to be conducted by the Wide Field Infrared Explorer (WIRE) and eventually by the Space Infrared Telescope Facility (SIRTF). WIRE, to be launched in February 1999, will be able to detect brown dwarfs as cool as 300 K, implying masses as low as 20 Jupiter masses. The combination of the ground- and space-based searches will lead to a full understanding of the lowest-mass objects that the star formation process can produce. The next decade of research will also lead to great advances in our knowledge of planets orbiting nearby stars. Two instruments will complete a census of planets around stars in the solar neighborhood. The four outriggers of the Keck Interferometer will measure the positions of stars with 10- to 20-microarc-sec accuracy, surveying 1,000 stars for planets as small as Uranus in a jovian orbit. The Space Interferometer Mission (SIM) will push this accuracy down by almost a factor of 10, enabling searches for planets of a few Earth masses around the closest stars and looking for Jupiter-mass planets around stars as far away as 1 kpc. These astrometric results will complement ground-based radial-velocity studies, which to date have been the most fruitful method for finding planets. Other indirect methods that may prove useful in determining the incidence of planets are microlensing and transit photometry. Both techniques are limited to studying distant stars. In a few cases it may be possible to detect the visible or near-infrared light from planets directly despite the glare of the parent star. Ground- or space-based telescopes equipped with coronagraphs using sophisticated wavefront control to reduce scattered and diffracted light might image planets directly around the closest stars. Whether the Next Generation Space Telescope (NGST) will be equipped with such instrumentation is at present unknown, but subsequent large telescopes could be. Finally, there is the ultimate goal of finding and characterizing other Earthlike planets. The mid-infrared is the natural wavelength band in which to carry out these investigations because of the improved contrast between the planet and its parent star, and because of atmospheric tracers such as H2O, CO2, O3, and CH4 that are found uniquely in the 7- to 17-μm band and that may serve as markers of habitable, or even of inhabited, planets. Progress is being made in developing the performance requirements for the Terrestrial Planet Finder (TPF) and for developing the key technologies needed for such an instrument.