Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects (1998)

The study of SMOs is currently in a renewed state of vigor after several decades of false starts and frustrations. The key to the new successes lies in advancement of ground-based telescopes and detector systems, buttressed by results from key spacecraft programs (e.g., the Infrared Astronomical Satellite, Hubble Space Telescope, Hipparcos, and Galileo) and theoretical studies of growing power and fidelity. The challenge for NASA and other funding agencies is to foster these programs in such a way that they contribute to NASA's ultimate programmatic goal of discovering terrestrial planets in orbit about other stars, but without encouraging a premature narrowing of focus toward a single, high-cost technique or mission.

Studies of SMOs have made significant progress in the last few years, but they are still in their infancy. The number of known brown dwarfs and giant extrasolar planets is still relatively small and provides an insufficient basis for drawing definitive conclusions about the range of properties exhibited by these objects. Moreover, detailed information about individual SMOs is generally lacking. Thus, the most compelling issues concerning SMOs as objects of intrinsic interest are the following:

• Devising detection strategies to increase the population of known SMOs beyond the several hundred expected from the DENIS and 2MASS surveys and thus,increase the extent of SMO parameter space accessible for study; and

• Performing spectroscopic and other diagnostic studies to characterize individual, nearby SMOs.

The exciting results from ongoing programs designed to detect SMOs lead to a natural desire to extend this work to fully exploit the ultimate capabilities of current detection techniques. Indeed, a natural evolution is already apparent. For the past decade there have been a number of limited radial-velocity and astrometric searches that utilized small- and moderate-aperture telescopes. The results of these surveys were sufficiently encouraging to stimulate the start of much larger and more ambitious surveys of even higher precision utilizing the first of the new generation of large-aperture telescopes. These new surveys are increasing the sample size by an order of magnitude.

The next phase in this evolution is to use space-based facilities to undertake measurements that cannot be done from the surface of Earth. The Space Interferometry Mission (SIM) and the Kepler photometric mission, both in the planning stages, represent efforts in this direction. This evolutionary path will lead to a steady increase in the extent of the SMO parameter space accessible for study. The confidence gained with the exciting results from each step helps to justify the greater complexity and cost of the next step.

This evolutionary process must be carefully fostered and nurtured. A balanced program that fully explores all of parameter space is essential. In particular, to advance SMO studies in the near term and to optimize the development of space-based facilities NASA must balance its major space expenditures with adequate funding for the early steps involving ground-based techniques and flight demonstrations. For example, a significant amount of effort needs to be delivered to ground-based, transit-photometry efforts to provide a dramatic demonstration of the viability of this technique. The detection of a new 51 Pegasi-like system by photometry would make a compelling case for future space-based photometry missions. Similarly, the vigorous exploitation of the capabilities of current ground-based interferometers will help researchers to improve the designs of much larger and more expensive future ground-and space-based interferometers.

Extensive astrometric surveys with facilities such as NPOI (Navy Prototype Optical Interferometer) PTI (Palomar Testbed Interferometer), and CHARA (Center for High Angular Resolution Astronomy) and others will also help train the next generation of scientists who will then design and fully utilize the much more expensive space missions such as SIM and Planet Finder.

The great difficulty in detecting SMOs is the result of their low luminosity. It is now clear, however, that the so-called stellar mass function (i.e., the number of objects in a given mass range) does not continue to rise through the bottom of the main sequence to the SMO domain. Rather, the mass function is flat or gently declining through the transition to SMOs, with perhaps a rise below 10 M_J. Thus, SMOs are at best as numerous as stars, and perhaps only half so abundant. As a result, SMOs cannot contribute more than 10% of the mass of the galactic neighborhood. Thus SMOs will always be elusive candidates for discovery and characterization regardless of the increased capability of telescopes and instruments, because they do not stand out in sheer numbers against the background of the low-mass stellar population. The diverse techniques of detection described in Chapter 1 will continue to provide cohorts of candidates whose true nature will be understood through more detailed observations with different tools.

The technique of choice for studying individual SMOs will continue to be spectra, the bulk of which will be collected from the ground. Continued improvement in optical and infrared detectors, the greater availability of large telescopes, and progress in developing adaptive-optics techniques (to separate the light of close companions from that of their parent star) are all required to probe the spectra of SMO candidates discovered to date and anticipated in the coming years.

These technologies, developed on ground-based systems, have direct application to the goal of acquiring spectra from terrestrial planets around nearby stars using space-based telescopes. However, the continued advancement of SMO studies requires that NASA encourage a range of approaches that will have broad scientific benefit for the detection and characterization of SMOs. This will have the additional

potential for alternative and unexpected solutions to the problem of characterizing extrasolar terrestrial planets.

Close-orbit SMO companions, such as 51 Pegasi B, may not be amenable to spectroscopic study in the foreseeable future. To go beyond the meager information available today on these objects will require bringing multiple techniques to bear on these systems. Improvements in astrometry may allow objects discovered by radial velocity to be detected as well by astrometry, which does not suffer from the sin i ambiguity in mass. Hipparcos data have already provided a partial constraint on one such system. Other novel approaches to exploring such systems, such as detecting the reflected light component from the companion, will need to be invented and refined as we move away from the novelty of discovery to the dilemma of characterizing the close-in companion SMOs.

Studies of SMOs besides improving understanding of these intrinsically interesting objects are likely to make significant contributions to a number of broader goals in astrophysics and the planetary sciences. SMOs are a bridge between stars and the giant planets of our solar system. The physics of their atmospheres and interiors represents a genuine transition from stellar physics to planetary physics, and the ability to model their evolution, structure, and appearance represents an important test of radiative transfer, thermodynamics, and condensed-matter physics over a broad range of physical parameter space. Moreover, their frequency among various stellar populations provides important constraints on the physics of star formation and the environments within which it occurs.

As stellar companions, the SMOs detected to date appear to be representatives of both the continuation of the star formation process to lower masses, and the continuation of planetary formation through the mass of Jupiter and above. The evident ability of SMOs to undergo orbital migration from their point of formation to almost the edge of the parent star challenges our understanding of disk-planet-star dynamics. Furthermore, the history of SMOs through their formation, migration, and long-term orbital states is a principal consideration in assessing the likelihood of the presence of terrestrial planets in a given system, as well as the small-body environment to which those planets have been subjected through time and the current detectability of such planets through the haze of a zodiacal dust population.

To highlight some of the important areas where SMOs are likely to make contributions to broader scientific goals above and beyond their contributions to improving current understanding of the general theory of star and planet formation, the steering group has selected three particularly compelling examples:

• Modeling of the atmospheres and interiors of SMOs;

• Testing models of the formation of SMOs; and

• Understanding the stability and evolution of multiplanet systems.

In the following discussion of these examples, areas and issues of particular importance are highlighted.

The study of brown dwarfs is in its infancy, with initial thermal profiles and the identification of several constituents in the upper atmosphere of only one specimen. Further efforts will lead to a better understanding of convection in planetary atmospheres, the structure of the deep jovian atmosphere, and the chemistry of cool stars and jovian atmospheres. The recent availability of calculations of the molecular-band structure of hot water was essential to the analysis of Gliese 229B, whose spectrum is dominated by water. Previous use of data on the water bands compiled for use in studies of Earth' s atmosphere failed to account for most of the observed absorption. Improvement in the interpretation of Gliese 229B's spectrum relies on calculations and measurements of hot molecular bands. Properties for methane are particularly important because this gas contributes significant opacity in atmospheres ranging from Jupiter's to that of 1000-K-brown dwarfs. Such calculations would provide, for example, a better interpretation of Gliese 229B's spectrum at 2 µm. Gliese 229B's flux at this wavelength is diagnostic of this brown dwarf's mass, which, at present, is highly uncertain. Laboratory and associated computational efforts to construct accurate spectral-line lists are essential to accurate modeling of substellar-mass objects.

Cloud formation is one example of atmospheric processes that are demonstrably non-uniform, as visual examination of Earth, Jupiter, and Saturn attests. Yet clouds play key roles in the atmospheric energy balance, surface brightness, and compositional variation of the atmosphere. The realistic incorporation of cloud formation into SMO models, as well as other processes that are spatially or temporally variable or not in local thermodynamic equilibrium (a standard current assumption), requires an increase in computational speed and capacity above that which is now used for atmospheric models. Parallel, vector, or superscalar processors must become widely available to take full advantage of current understanding of the physics involved in modeling these objects.

This situation is entirely analogous to the transition, over the past 5 years, from gray (frequency-averaged) to non-gray (frequency-dependent) model atmospheres: ¹ while it was long recognized that the simplifications of gray models could lead to results potentially misleading to observers, only the advent of widespread, high-speed computational capability enabled accurate, non-gray models to be implemented. The steering group emphasizes the term “widespread” because often in the past the limited advent of new computational capabilities has led to oversubscription, and some disciplines have failed to benefit rapidly from the new tools. The advent of non-gray models led to the identification of unexpected flux enhancements in certain wavelength regions, which have provided important guides to observers.

Additional observational requirements for atmospheric physics are similar to those in every field in astronomy: larger telescopes and more sensitive detectors. If improved, the spectral resolution and signal-to-noise ratio of observations would enable investigations of stratospheric physics. Emission lines in the stratospheres of planets tend to be narrower than the tropospheric lines and therefore amenable to high-resolution spectroscopy. These lines have yielded information on the chemistry and thermal profile of planetary stratospheres, signatures of photochemistry, and stratospheric dynamics. Interest in these subjects has recently been ignited by the discovery that Neptune experiences photochemistry, possibly driven by ultraviolet emissions by the interstellar medium. In addition, Galileo's entry probe revealed a thermal profile indicating that Jupiter's stratosphere is heated by gravity waves.

Although the observable properties of SMOs are not as sensitively dependent on the details of the deep interior as they are on the atmosphere, a proper treatment of the interior physics is nonetheless important as observations better constrain the physical radii and surface gravities of SMOs. The complex progression from highly degenerate objects of great age or low mass, to the more thermally expanded youthful or massive SMOs, demands a thorough understanding of the equation of state of hydrogen and helium to gigabar pressures and higher.

Experimental and theoretical work in this area is reasonably vigorous at present, largely through the efforts of the international high-pressure physics community to organize themselves through frequent conferences and workshops. Continued support of, and increased capabilities in, the experimental and theoretical investigations of the behavior of materials at high pressures are essential to accurate modeling of substellar objects. One outstanding issue directly tied to high-pressure physics, for example, concerns the phase separation of minor species from the hydrogen pressure-ionized phases. Separation of this type leads to the disappearance of such species from the observable atmosphere in well-mixed SMOs, and potentially leads to ambiguities in the interpretation of the presence or absence of certain features in the spectra of SMOs.

Over the past four decades, planetary sciences have identified and investigated the key physics and chemistry that direct the evolution and determine the characteristics of a planet. Yet each planet in our solar system has introduced new considerations and effectively modified our understanding of the others. Brown dwarfs offer an opportunity to investigate a range of planetary bodies in new mass regimes. Detection techniques that identify systems close enough for spectroscopic analysis (e.g., the DENIS survey) are therefore extremely useful. The complexity of planetary atmospheres requires many examples for its understanding.

As is the case for the modeling of SMO atmospheres and interiors, simulations of the formation of SMOs rely on sophisticated theory enabled by increasingly powerful computational tools. Hydrodynamic and magnetohydrodynamic simulations of disk processes, including fragmentation and two-stage planet formation, are computationally intensive and benefit from insights leading to more efficient and accurate protocols for solving the equations.

Formation studies rely on the stimulus of observations perhaps to a greater extent than do the structure models. A broad range of observational strategies for detection (i.e., astrometry, photometry, radial-velocity, and microlensing studies) and characterization (e.g., spectroscopic studies over a broad range of the electromagnetic spectrum) are required to provide the data essential for constraining formation mechanisms. For example, the paradigm of giant planet formation prior to the discovery of 51 Pegasi B was that such objects should not be found very far inward of the disk's so-called ice-line (i.e., the minimum radial distance at which ice can condense and, hence, contribute large amounts of solid to planet formation). Although gap formation and inward migration of a giant planet in a gaseous disk were considered, there were no predictions that such a process could lead to a significant number of systems with very close-in giant planets. With the discovery of 51 Pegasi B, the paradigm was forced to shift, and a substantial effort was made in developing multiple mechanisms for moving giant planets inward, and stopping them. This, in turn, has led to broader thinking on the possible architectures of planetary systems.

Improvements in observed spectra of SMOs and their stellar primaries may play a role in deciding among formation mechanisms. The planetlike process requires addition of large amounts of heavy elements in the form of accreted grains, so that the metallicity of the

companion should be significantly greater than that of the primary —as is the case for Jupiter relative to the Sun. Although direct formation by the starlike process should lead to a metallicity comparable to that of the primary, subsequent evolution (e.g., orbital migration) could increase the metallicity of the secondary, so that this test would need to be applied with caution. Also a challenge will be accounting for all of a particular element that may exist in a large number of molecular forms in a typical SMO's spectrum, as well as accounting for condensation and phase separation processes (e.g., the depletion of water in Jupiter's atmosphere2 and helium in Saturn's). Despite these complications, the aspiration that detailed spectra of SMOs could constrain formation processes will continue to be an important driver of improved observational spectroscopic techniques and more detailed synthetic spectra.

An important frontier that remains is the issue of stability of systems containing multiple planets. This, too, is a computationally intensive undertaking that must confront nonlinear dynamical systems with multiple degrees of freedom and, hence, chaos. Studies of the stability of various planetary configurations stretch back over many decades. But only in the last several years have they been informed by the discoveries of jovian-mass SMOs.

In this regard it is important to understand the effects of giant planets, in various orbital and mass combinations, on the population of small bodies and dust in a planetary system. The solar system 's giant planets are arrayed in such a way that they have efficiently ejected leftover planetesimals into distant orbits (e.g., the Oort Cloud and Kuiper Belt) early in the history of the solar system, and later helped to bring perturbed orbits inward to produce copious quantities of short- and long-period comets. Both sets of events have been key in determining the habitability of Earth's environment and its capacity for catastrophic change through time, and may have played a still-poorly understood role in fostering life and its evolution.

Different configurations of giant planets will produce different evolutions of small-body populations in other planetary systems, with perhaps different results for the habitability of terrestrial planets in those systems. Furthermore, the dust loading in other planetary systems, as affected by giant-planet perturbations on the cometary supply sources, has important implications for the detectability of terrestrial planets using the approach of space-based optical/infrared interferometry.

Further progress in understanding the origin and stability of planetary systems will come from the wider availability of increased computational power, the study of analogs in our own solar system (e.g., the moons of Uranus), and continued discovery of new planetary systems. Key to the flourishing of these activities will be NASA's recognition of their important intellectual and technological connection to the search for terrestrial planets around other stars.

The study of SMOs requires communication and collaboration across a broad range of disciplines in the physical sciences. Formation models require application of hydrodynamics, magnetohydrodynamics, and other disciplines hitherto brought to bear on the star-formation problem. The properties of individual brown dwarfs are modeled through interdisciplinary

efforts among the atmospheric sciences (including cloud microphysics), coupled with atomic physics, molecular equilibrium thermodynamics and reaction kinetics, radiative transfer, and hydrodynamics, as well as high-pressure physics (experimental and theoretical) for the interiors. Observational techniques require fundamental laboratory work in atomic spectroscopy (for calibrating radial-velocity studies, for example), as well as molecular spectroscopy, high-precision photometry and astrometry, and coronographic imaging and spectroscopy. Particularly noteworthy is the confluence of techniques from the planetary sciences as well as astrophysics, as modelers attempt to understand the nature of brown dwarfs by applying tools derived for planetary atmospheres as well as those for stellar atmospheres.

Studies of SMOs have direct relevance to a number of long-standing scientific goals and priorities. The most obvious of these is the role of SMOs as a testing ground for honing the instrumentation and techniques necessary to detect extrasolar terrestrial planets. Another possible area is the contribution of SMO studies to the identification of missing mass in the universe. Although it appears that SMOs do not constitute the bulk of the matter in the universe, they do represent unique probes of galactic structure. Observations of microlensing events, a technique developed originally to search for missing mass in the form of massive compact halo objects, have particular promise for probing the mass function of brown dwarfs and understanding the composition of the galactic halo. With appropriate development, microlensing may provide a shortcut to the detection of extrasolar terrestrial planets.

Solid programs are already in place to measure the brown dwarf mass function (MF). Photometric determinations of the MF are proceeding well. For example, the 2MASS and DENIS surveys will probably find several hundred solar-neighborhood brown dwarfs. HST/NICMOS observing time has already been awarded to measure the bulge luminosity function (LF) in the infrared, which is likely to extend the bulge MF measurement from the present 0.3 solar masses to the hydrogen-burning limit. HST time for additional studies of clusters, including proper-motion studies to eliminate field contamination, is likely to be awarded through the normal competitive process. In addition, ongoing radial-velocity searches for extremely low-mass (planetary) companions of nearby stars will automatically probe the brown dwarf regime. Continued support for programs to measure the mass function of brown dwarfs is appropriate.

By contrast, ongoing microlensing surveys are at a critical juncture and will require significant attention if they are to fulfill their potential for probing the brown dwarf MF as well as settling the composition of the halo. Observations reveal that too many short microlensing events are seen toward the bulge compared to what is expected from the photometrically determined bulge MF and simple models of bulge kinematics. The two most plausible explanations for this discrepancy are that the bulge MF turns up at low masses or that bulge kinematics are more complicated than the simplest models. It is probably not possible to distinguish between these two possibilities with standard microlensing searches because the one observable microlensing parameter, the event time scale, is proportional to the square root of the

mass divided by the transverse speed. Hence, it cannot distinguish between MF and kinematic effects. To break this degeneracy, observers must measure higher-order microlensing effects.

From the standpoint of measuring the brown dwarf MF, the most important such effect is the deviation of the microlensing light curve that occurs when the lens transits the source. This occurs for about 13% of brown dwarf events with giant-star sources, or perhaps eight events per year. The perturbation lasts about a day and, if it is accurately measured, breaks the degeneracy of the time scale into two quantities: the proper motion of the lens relative to the source and the product of the lens mass and lens-source distance. Measurement of this effect in a few dozen events would greatly clarify the puzzle of the short microlensing events observed in the bulge.

The hardware and software requirements to make these measurements are almost identical to the requirements for planet detection. In particular, the duration of the perturbation is the same as that expected for Jupiter-mass planets. Two worldwide networks have already been established to carry out the necessary observations, and much of the required equipment is in place or soon will be. Both networks have shown that they are capable of obtaining the very accurate and nearly continuous light curves needed to make these measurements. However, these are large-scale integrated observing operations, and as they ramp up to “production mode,” it is possible that new bottlenecks will be identified in the hardware and/or data analysis. Relatively modest assistance could have a major impact on the productivity of efforts to measure the brown dwarf mass function via observations of microlensing events.

A potentially major structural problem could have a damaging impact on microlensing observations: the two largest surveys (MACHO and EROS) are carried out primarily to search for dark matter and are funded primarily by agencies that see this search as part of their mission (e.g., the Department of Energy, the National Science Foundation, and France's Atomic Energy Commission). The follow-up observations, whose primary objectives are the search for planets and brown dwarfs, and the study of galactic structure, are funded in the United States by NSF and NASA and in other countries by parallel organizations.

The follow-up observations, of course, require the primary searches to provide events to follow, but the follow-up, although useful to the interpretation of the primary searches, is not critical to it. Thus, left to their own, the primary searches could well suspend operations while they initiate new, more ambitious search programs. Indeed MACHO will do exactly this on January 1,2000. Such a hiatus would be extremely disruptive to the follow-up networks that have obtained long-term access to telescope time in exchange for various commitments. The various agencies supporting primary and follow-up microlensing observations should work together to minimize potential disruptions caused by differences in their primary goals.

Measurement of higher-order microlensing effects is also required to determine the nature of the lenses toward the LMC. Although there are several ideas for space missions that would definitively resolve this question (Space Interferometry Mission, or a microlens parallax satellite), it is also possible to make substantial immediate progress from the ground. If the lenses lie in the halo, then the reflex motion from Earth's orbit should give rise to 1% parallax asymmetries in the light curves. These could be detected by follow-up observations similar to those now under way toward the bulge, but smaller in scope by a factor of four. Modest support to initiate a search for parallax asymmetries as an adjunct to the microlensing program will yield additional important information on the nature of the objects creating the lensing events.

NASA's current programmatic goal of discovering another Earth is a laudable one upon which the steering group does not seek to lay specific recommendations. In addressing this goal, however, NASA should do the following:

Continually assess the new information that studies of SMOs are providing on the formation, frequency, and characteristics of planetary systems, and invest judiciously in developing observational and theoretical techniques that will foster new discoveries. This investment should be in addition to the funding NASA is already providing for technological development of future large projects such as the Space Interferometry Mission and the Terrestrial Planet Finder. The funding must be flexible and peer-reviewed in recognition of the nature of the activities, which are distributed principal-investigator-based projects to observe and model SMOs using a variety of different approaches. The small-scale nature of these activities suggests that existing procedures (e.g., periodic peer review of proposals and resulting publications) will be adequate to identify and prioritize the approaches and techniques deserving of additional investment.

Invest with care in select ground-based facilities, instruments, and computational programs that will significantly broaden the near-term opportunity for innovation in the identification and characterization of SMOs. Addressing the broader issue of the appropriate balance between the support of ground-based programs by NASA, the NSF, and other appropriate agencies is beyond the scope of this report. This important topic is best addressed by the decadal survey committee in the context of the findings of the study on the federal funding of astronomical research currently being conducted by the NRC's Committee on Astronomy and Astrophysics.

Consult with other agencies (e.g., the National Science Foundation) to avoid duplication and to open a broader set of opportunities for research and discovery through cooperative or collaborative funding.

In sum, research on SMOs is at the heart of trying to understand the matter content of the universe, the ubiquity and properties of planetary systems, and the relationship (in both genesis and physical properties) between stars and objects not massive enough to ever become stars. By studying SMOs we extend our understanding of the cosmos from the ubiquitous macroscale of stars through to the planets and, hence, ever closer to the human realm.