ÿþ



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 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects 2 Observational Characterization of Nearby SMOs CURRENT STATE OF OBSERVATIONS Abundance of SMOs The ensemble of candidate SMOs detected by various means numbers in excess of 80 objects, although the identification of most of these as bona fide SMOs is questionable (see presentation by G. Basri in this chapter). Most of the 12 candidates detected by indirect means (i.e., radial-velocity techniques) as of the workshop are almost certainly substellar, although we know strictly only M sin i (the mass multiplied by the sine of the orbital inclination relative to the line of sight to Earth). Five objects are seen directly as companions orbiting other stars; one of these (Gliese 229B) shows methane (CH4) in its spectrum, which, coupled with other constraints, requires the visible atmosphere to be at about 1000 K (see presentation by M.S. Marley in this chapter). Together with the age of the parent star, estimated from stellar evolution models, this tightly constrains Gliese 229B as an SMO with a mass possibly as low as 30 MJ. Of the remaining four objects, two, detected directly as companions, likely are at the bottom of the stellar main sequence but are not substellar. The remaining two form a binary system, and lithium is seen in the spectrum of both. Lithium is an important diagnostic element because it is destroyed by fusion reactions in the interiors of stars and objects just slightly below the hydrogen-burning mass limit (see presentation by Both globular clusters (see presentation by I. King in this chapter) and star-forming regions (see presentation by L.A. Hillenbrand in this chapter) are important environments in which to search for brown dwarfs. In young stellar clusters, where even low-mass objects are bright and hence relatively easy to detect, eight candidate SMOs have been discovered. All but one show lithium in their spectra, and although they belong to young clusters they are evolved enough to have destroyed lithium if they were stellar in mass. By contrast, three dozen candidate SMOs in star-forming regions, identified by their red color, are too young to have destroyed their lithium regardless of their mass, and hence their candidacies are highly tentative. Nine  free-floaters, i.e., objects not in orbit around other stars and not associated with clusters or star-forming regions, are by various lines of evidence considered to be candidate SMOs. The detailed description given above illustrates well the complexity of the situation with respect to identifying SMOs and determining the abundance of such objects in the galactic neighborhood near the Sun. Clearly objects are now being detected, thanks to increasingly sensitive detectors, larger telescopes, more capable computers for processing data, and novel

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects observational techniques buttressed by detailed theoretical models. However, only a handful of the direct detections are firmly established as true SMOs, and the only knowledge researchers possess of the indirect candidates are the lower limits to their masses. Recent data from the Hipparcos satellite on one of the systems has, however, set an upper limit to the inclination which firmly establishes the candidate as substellar in mass. The number of candidates detected so far in the regions of the sky searched suggests that, extrapolated to the entire sky, thousands more are detectable with present instruments. But the detailed study of such objects, including the collection and analysis of spectra, awaits increased sensitivity of detectors and greater availability of large telescopes. Much like Moses, observers are at present restricted to gazing from afar upon—but not yet experiencing—a promising future rich in the detection and study of SMOs. Spectroscopic Studies of SMOs Spectral analysis of SMOs is critical not only to understand the physical properties of these objects but also to identify molecules that tightly constrain the atmospheric temperatures and, hence, allow firm assessment of the objects' masses. Key molecules in this regard are methane (CH4) and ammonia (NH3), which become increasingly abundant at cooler atmospheric temperatures at the expense of carbon monoxide (CO) and molecular nitrogen (N 2). The challenge will be to detect small amounts of methane in objects that are close to the edge of the stellar main sequence: spectra extending further into the infrared will be helpful in this regard but challenge current capabilities. Sensitive upper limits on molecular species that are present in cool stars but should be condensed out of the observable atmospheres of lower-mass and hence cooler objects is another spectroscopic test of SMOs membership that further emphasizes the need for high sensitivity. The direct detection and analysis of spectra of the most massive SMOs have proceeded sufficiently that many observers are endorsing the creation of an additional letter in the traditional stellar classification sequence of main sequence objects: O, B, A, F, G, K, M. Objects near the main sequence edge have spectra qualitatively different from M dwarfs, in that titanium oxide (TiO) and vanadium oxide (VO) are absent, and other molecular lines dramatically perturb the spectrum from blackbody. A proposal to designate SMOs as “L” dwarfs is not simply an exercise in nomenclature but a recognition of the discretely different nature of the spectrum of objects with surface temperatures of 1500 K and lower.

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects THE PHYSICAL AND ORBITAL CHARACTERISTICS OF KNOWN SUBSTELLAR-MASS OBJECTS Gibor Basri University, of California, Berkeley This review will serve largely as a census of the observationally confirmed and likely brown dwarfs discovered to date. Their physical characteristics have, for the most part, not been measured in detail. They generally conform to our theoretical expectations, which are summarized in other presentations. Turning first to brown dwarfs in binary systems, the most obvious thing to do is simply look near a star. This approach yielded what stood for almost a decade as the most intriguing brown dwarf candidate, GD 165B, as well as what now stands as the most incontrovertible brown dwarf: Gliese 229B. The former object is unfortunately near the minimum possible main sequence temperature (1700 K, though this remains uncertain), while the latter is well below it at around 1000 K. Gliese 229B is probably 2 to 5 Gy old, with a mass of 30 to 50 Jupiters. Its orbit is too wide (40 AU) to have been measured yet, although eventually we should be able to get a dynamical mass. It is the only substellar companion found in a survey of ~300 M stars. The success rate of imaging surveys near white dwarfs, and in clusters, is similar. Most of the brown dwarfs known in binary systems (~10 out of about 700 targets) have been found through radial-velocity work. One cannot be sure of the nature of any given one (because their orbital inclinations are unknown), but can be sure that most of them are really substellar objects because of the mass-limit distribution of stellar companions. Their orbital parameters are biased to small separations because of the search technique, but they have an eccentricity distribution similar to that of stars, and different from that of the extrasolar planets. Thus it is reasonable to imagine that they form like stars. Very recently, one system involving a binary brown-dwarf pair has been found. These massive brown dwarfs are quite close (0.03 AU) to each other in a moderately eccentric orbit. The mass distribution of brown dwarf companions appears to continue that for low-mass stars. Below 10 Jupiter masses there is a sudden rise in the number of objects per primary, despite the increased difficulty of detecting them. The most fruitful search for brown dwarfs so far has been in young clusters, particularly the Pleiades. Young brown dwarfs are brighter, but must be distinguished from late M stars (which they very closely resemble). The “lithium test” has proven a powerful means of doing this: all directly observed brown dwarfs, except Gliese 229B, currently gain their pedigree this way. Brown dwarfs (~15) have been found in the Pleiades from the substellar limit down to about 30 Jupiter masses. This allows the most direct measure of the substellar initial mass function (IMF) to date, which indicates a mass function rising as the reciprocal of mass to very low masses. Preliminary results in the Trapezium cluster may show a turnover at low-mass stars; this relies on evolutionary calculations to yield mass. It is important to extend the work to other clusters; we do not know how universal the substellar IMF is. Finally, free-floating field brown dwarfs have recently been confirmed. Preliminary results from the DENIS and 2MASS all-sky infrared surveys indicate a space density compatible with that predicted from the Pleiades. Objects are found that are clearly cooler than M stars —a new spectral class “L.” Their spectra provide good evidence that dust is an important atmospheric constituent.

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects GROUND- AND SPACE-BASED SPECTROSCOPY OF THE COMPOSITION OF BROWN DWARFS AND EXTRASOLAR PLANETS Mark S. Marley New Mexico State University The molecules found in the atmosphere of a brown dwarf or extrasolar planet constrain atmospheric thermal structure, dynamics, and chemistry. In this review I focus on what we can learn from the spectroscopic detection of molecules in these atmospheres and illustrate with examples from Gliese 229 B and the solar system's jovian planets. The composition of a static planetary atmosphere, with no external perturbation, tends toward local thermodynamic chemical equilibrium. Thus at the relatively low temperatures (~100 to 200 K) and high pressures of the solar system's jovian planets, C is found in CH4 (methane), O in H2O, and N in NH3 (ammonia). At higher temperatures thermochemistry favors CO, H2O, and N2. Indeed it was the spectroscopic detection of CH4 in Gliese 229 B that incontrovertibly confirmed the identity of the object as a brown dwarf. Ammonia is predicted to also be present in the upper atmosphere of the brown dwarf, but it has yet to be detected. Hence, to first order, the presence or absence of CH4 and NH3 can serve as thermometers of atmospheric temperatures. Methane is easily detectable in the near infrared. Ammonia is detectable in the visible and near 10 μm. A major difficulty, however, is that there are essentially no data on the opacity of these molecules at temperatures greater than the “traditional” planetary temperatures. Departures of atmospheric composition from equilibrium are especially interesting. Molecules such as CO, PH3, GeH4, and AsH3 have all been detected in Jupiter's atmosphere at abundances many orders of magnitude higher than expected from equilibrium chemistry. The presence of these non-equilibrium molecules is taken to be evidence of convection. Since convective time scales are shorter than chemical equilibrium time scales, these molecules can be dredged up from deep in Jupiter's interior and transported to the visible atmosphere. Likewise, the detection of CO at Neptune but not at Uranus apparently provides information on the relative vigor of convection in these two atmospheres. At Gliese 229B the detection of CO in abundances far in excess of that predicted for chemical equilibrium clearly requires that the visible atmosphere (near 800 to 1400 K) must also be convective. Yet many atmosphere models find that the radiative-convective boundary lies far deeper, below 1700 K. However, some models predict an additional, detached, upper convection zone, which is consistent with the CO detection. The presence of Cs at Gliese 229B also requires convective transport as this element should not otherwise be present. However, the lack of TiO and other refractory diatomic species suggests that the atmosphere is not fully convective to the depth (below 2000 K) where these molecules condense. Taken together, these results support the presence of a detached convection zone. Thus CO and Cs may be tracing the vertical convective structure of the brown dwarf while strengthening the earlier suggestion that a radiative region also exists far below Jupiter's cloud tops. PH3 is also potentially detectable at Gliese 229B by space-based platforms and will provide an important additional test. Measurements of the abundances of this suite of molecules in a variety of objects will map out the atmospheric dynamics of substellar objects. Incident radiation can also produce important non-equilibrium species. Thus many photochemically produced hydrocarbons are found in the atmospheres of the solar system's jovian planets, including C2H2 and C2H6, that would not otherwise be expected. These relatively fragile molecules escape destruction because there is little vertical mixing in the stratospheres where these molecules form. A rich variety of photochemical products will likely be found in

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects the atmospheres of the extrasolar planets, particularly those with warm atmospheres and large incident fluxes. Hazes, produced by the condensation of some species, can produce signatures in the spectra of these objects far in excess of what might be expected given their small mixing ratios. Condensation of major species also will play an important role in the spectra of substellar objects. The near-infrared spectra of Gliese 229B is best fit by cloud-free models, yet the depressed visible flux implies the presence of grains. If there is indeed a radiative region above the condensation region for silicate and iron grains, as the molecular data suggest, it is likely that the Gliese 229B grains are produced by low-abundance species (sulfates?). Mie scattering theory suggests that submicron grains have the requisite scattering and absorption properties to preferentially affect the visible, but not the infrared, flux. However, a self-consistent chemical equilibrium, radiative transfer, grain condensation scenario has not yet been worked out. The extensive study of these very processes in planetary atmospheres will guide further investigation. We clearly must bring to bear the expertise of both planetary scientists and astrophysicists to correctly interpret the spectra of extrasolar planets and brown dwarfs. History tells us that interactions between such formerly disparate fields can be difficult at times, but the ultimate payoff will be a rich and exciting new science. COMPARATIVE SPECTROSCOPY OF BROWN DWARFS AND VERY LATE MAIN SEQUENCE STARS Rafael Rebolo Instituto de Astrofisica de Canarias To tell a brown dwarf from a very-late-type dwarf star was, until very recently, an extremely complex task. To the renowned difficulty of measuring masses with sufficient precision, one had to add the lack of reliable spectroscopic substellarity criteria. At present, however, at least two of those criteria have proved to be useful, contributing in a decisive manner to astronomers revealing the nature of a good number of brown dwarf candidates. On the one hand, the substellar detection of near-infrared methane bands, first suggested by Tsuji et al. as a substellar indicator,1 has confirmed Gliese 229B as a cool brown dwarf (effective temperature, ~1000 K), and detailed spectroscopic studies on this object currently serve as a guide to both theoretical modeling of very cool atmospheres and the design of observational strategies aimed at detecting similar objects. The spectral differences of Gliese 229B with respect to the least massive stars are huge and rule out the possibility of misinterpretation or uncertain identification of similar objects. On the other hand, the spectra of young brown dwarfs are expected to display features very similar to those of the least massive stars. We are fortunate enough, however, to have a criterion with which to identify these brown dwarfs by means of a simple spectroscopic test based in detecting lithium lines in their spectra—particularly the resonance lines at 6708.8 1   T. Tsuji, K. Ohnaka, and W. Aoki, “Spectra and Colours of Brown Dwarfs,” in The Bottom of the Main Sequence—And Beyond, C.G. Tinney (ed.), ESO Astrophysics Symposia, Springer-Verlag, Berlin, 1995, p. 45.

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects angstroms.2 Lithium is a fragile element, efficiently destroyed in fully convective stars, whereas it is preserved in brown dwarfs with less than 65 Jupiter masses. The test provided the first successful results on objects in the Pleiades cluster3.4 where, thanks to the detection of lithium, we are now able to determine the exact location of the substellar boundary, with utmost precision in terms of luminosity and spectral type. At the age of the Pleiades, the minimum mass for lithium burning coincides with the minimum mass for hydrogen burning. A substellar sequence in the Pleiades has been found down to spectral types later than M10, which appears to correspond to a mass close to 30 Jupiter masses. In addition, a reasonably precise determination of the mass fúnction in the cluster shows an increase in the number of objects up to nearly 40 Jupiter masses. Other stellar clusters are currently under scrutiny (α Persei, Praesepe, Hyades, Orion, Taurus, etc.), and we expect that comparable information will be obtained in terms of quality and quantity to that currently available for the Pleiades. Finally, a very rapid succession of lithium detections is taking place for very cool, free-floating objects discovered in the course of large-scale infrared surveys such as the Deep Near-Infrared Survey of the Southern Sky (DENIS) or during proper-motion studies. There are currently three objects cooler than M9 that have passed the lithium test and are therefore brown dwarfs. It is very likely that they are intermediate-age brown dwarfs, with masses around 40 to 60 Jupiter masses, that have not had enough time to cool down to temperatures such as that of Gliese 229B. Their main observational properties will be compared with those of the brown dwarfs in the Pleiades and very late field stars. ESTIMATING THE ROLE OF BROWN DWARFS IN GLOBULAR CLUSTERS Ivan King University of California, Berkeley Brown dwarfs are probably present in globular clusters, but not in such numbers as to be important to the structure of the cluster and be detectable by their gravitational influence. Extrapolation of the observed mass functions of the lowest-mass visible stars suggests that a substantial fraction of the stars in a cluster could be brown dwarfs, but they probably contain less than 10% of the mass of a cluster. Since there is no reason to believe otherwise, we can assume that the mass function (MF) of a globular cluster extends beyond the hydrogen burning limit, into the brown dwarf domain. Thus our task is to determine globular-cluster MFs down to as small a stellar mass as possible; we can then extrapolate to smaller masses to get an estimate of the number of brown dwarfs. The first step in determining the MF of a globular cluster is to determine its luminosity function (LF). Since this is usually done in only a small part of the cluster, and stars of different 2   R. Rebolo, E.L. Martin, and A. Magazzu, “Spectroscopy of a Brown Dwarf Candidate in the Alpha Persei Open Cluster,” Astrophysical Journal, 389: L83, 1992. 3   G. Basri, G.W. Marcy, and J.R. Graham, “Lithium in Brown Dwarf Candidates: The Mass and Age of the Faintest Pleiades,” Astrophysical Journal, 458: 600, 1996. 4   R. Rebolo, E.L. Martin, G. Basri, G.W. Marcy, and M.R. Zapetero-Osorio, “Brown Dwarfs in the Pleiades Confirmed by the Lithium Test,” Astrophysical Journal, 469: L53, 1996.

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects mass have different spatial distributions, it is necessary to model the cluster dynamically in order to convert this local LF into a global LF for the cluster. The next step is to convert the LF into an MF, by means of a mass-luminosity relation (MLR). The result is quite sensitive to the latter, because the conversion factor is the derivative of the MLR. Since there are no empirical MLRs for low metal abundances, it is necessary to use a theoretical MLR. These are still uncertain, especially at the low-mass end, where stars are fully convective and their hard-to-calculate atmospheres control their structure. Because all globular clusters have large distance moduli, observing their faint stars requires going to very faint magnitudes. So far, the Hubble Space Telescope has been the best instrument for such work. MFs that reach below 0.15 solar masses have now been determined for 10 globulars, and in one of these the MF extends beyond 0.1 solar masses. If the low-mass ends (below 0.25 solar masses) are crudely approximated by dN/dm = cm−α, the values of α range from 0 to 1.0. There is reason to believe that the clusters with lower values of α have selectively lost low-mass stars through dynamical causes and that a pristine globular has α between 0.6 and 1.0. If we naively assume a power-law mass function from 0.01 to 0.8 solar masses, then the fraction of the total stars whose masses lie between 0.01 and 0.09 solar masses goes from 25% to 50% as α goes from 0.5 to 1.0, while the fraction of the cluster mass that is contained in those stars goes from 0.036 to 0.101. The data presented here do not differ greatly from the faint end of the MF found by Gould et al.5 for the galactic disk (when the latter is converted to refer to m rather than log m). They do not confirm at all, however, much larger values of α that have been reported in the literature. BROWN DWARF CANDIDATES IN STAR-FORMING REGIONS Lynne A. Hillenbrand University of California, Berkeley Understanding the origin of stellar masses and of the initial mass function remains one of the primary goals in star formation studies. Is the production of a stellar- or substellar-mass object mostly a self-regulating process (i.e., controlled by the interplay of mass accretion and outflow), or do environmental conditions (e.g., molecular cloud properties) play a critical role? Why do some stars, primarily those of low mass, form in relative isolation from their nearest neighbors whereas most stars—especially those of high mass—form in dense clusters? Does the detailed spectrum of stellar masses produced during the cloud fragmentation process differ from region to region, or is it “universal”? Of particular interest is determination of the mass spectrum near to and across the stellar/substellar boundary. Does the mass spectrum turn over, implying that nature prefers to make hydrogen-burning objects, or does it continue to rise, implying a significant amount of mass, at the current age of the galaxy, in extremely low-luminosity objects? Because substellar-mass objects are significantly more luminous and hotter at young ages (#DXLT# 10 My) than at more evolved stages (#DXGT#100 My), star-forming regions are ideal places to 5   A. Gould. J.N. Bahcall, and C. Flynn, “M Dwarfs from Hubble Space Telescope Star Counts. III. The Groth Strip,” Astrophysical Journal, 482: 913, 1997.

OCR for page 19
Failed Stars and Super Planets: A Report Based on the January 1998 Workshop on Substellar-Mass Objects search for them. Indeed, photometric and spectroscopic investigations of the stellar populations in nearby star-forming regions are now sensitive to the detection of objects with masses below the hydrogen burning limit. Deep optical surveys (in the R and I bands) have enabled study of young stars that are still partially embedded in their nascent molecular clouds. When optical photometry is combined with optical spectroscopy, stars can be individually de-reddened and located in conventional Hertzsprung-Russell (H-R) diagrams, from which stellar masses and ages can be derived, leading to the construction of stellar mass and age distributions. More recently, by combining infrared photometry (in the J, H, and K bands) with infrared spectroscopy, this traditionally optical technique has been successfully employed in regions where most of the stellar population is fully embedded in molecular cloud material. We are thus able to determine stellar masses and ages for stars that are obscured by 10 to 50 magnitudes of interstellar and local circumstellar extinction. Probing into the birthplaces of stars in this way means that we can measure stellar-/substellar-mass distributions for temporally and spatially coherent populations that are unaffected by evolutionary processes. Moreover, rich, dense, extremely young clusters permit identification of complete samples of stellar as well as substellar objects, avoiding the membership ambiguities associated with their identification and study in older open clusters. Results to date indicate that while several very good substellar candidates do exist in young star-forming regions, the relative numbers of these objects imply a mass spectrum that does not increase from the stellar- into the substellar-mass regime. However, studies of the stellar populations in star-forming regions are hampered by the effects of high extinction, nebular contamination, source crowding, and circumstellar emission. Thus great care is needed in data analysis and interpretation. Furthermore, the translation from observational quantities (e.g., colors and spectral types) to physical quantities (e.g., masses and ages) depends on accurate understanding both of the intrinsic properties of late-type (M6.5 to M9) stars and of theoretical predictions for pre-main sequence evolution. At present, large uncertainties remain in stellar colors, bolometric corrections, and temperatures (on the observational side) and in opacities, convection, and the effects of accretion (on the theoretical side). Future observations aimed at narrowing these uncertainties are required in order to make additional progress on the stellar-/substellar-mass distribution in star-forming regions, for example, with large, near-infrared spectroscopic surveys.