Brown Dwarf

Brown Dwarfs are an unusual and exciting class of star/planet that exists on the fine line that separates objects with sufficient mass to become a star, and those objects whose mass falls shy of the density necessary to start the fusion process. They are heavier than planets but not massive enough to trigger the nuclear burning of hydrogen and other elements which powers normal stars. They are, nevertheless, heated during their formation by gravitational contraction but then continuously cool as this energy is radiated away. Sometimes called Proto-planets by planetary scientists, these objects are also known to astrophysicists as Protostars.

Even among the brown dwarfs, there are several different types, all of which exhibit their own unique characteristics. Some of the observations below give one an idea of the different processes that operate within a brown dwarf. One can see the similarities between these objects and protoplanets in their development from Proplyds (proto-planetary disks) into proto-planets/stars, and then planets/stars - as well as their sensitivity to equilibrium chemistry. However, it is obvious that planets consist of heavier elements than are typically found in a Zero-Age Main Sequence star that follows a similar path of development, which separates at the minimum fuel and mass requirement to become a star.

As MASS is very important in stellar development, more detail should be accorded this subject. I previously mentioned mass playing a major part in every star or planets development, and so it is. Another important factor is the amount of hydrogen, helium, and Lithium a star has. These lighter elements are relatively easy to fuse, and thus promulgate a stars development. If these constituents are of heavier elements, they form "peculiar" stars, like the Wolf-Rayet stars, or do not start their fusion engines at all.

The amount of the lighter elements relates to the mass of the star in that the more there is of a particular (light) element, the more pressure there is at the core - where fusion begins. It is because of the fact that so many molecules are vying for the same volume of space that fusion begins, and the fact that they are easily "excited" - bounced to the next energy level by collisions between particles - also makes the fusion process easier to initiate. In a Brown Dwarf, the fusion fires are:

a) not engaged, and the light and heat observed is from fission or some less efficient energy liberation/radiation process,

c) the material is not being fused at all - the heat and light given off by these proto-planets/stars is caused by some internal geological process, and/or

d) we are witnessing the repercussion from accretion of a great quantity of atomized high molecular-weight material from an unobservable companion into the fusion envelope of a pre-ZAMS star.

Using current theoretical evolutionary tracks and adopting an age interval of 1-5 Myr (Million years) for S Ori 47 of the s Orionis cluster, it is estimated that the mass of S Ori 47 is 0.015+/-0.005 M_solar (the minimum mass for deuterium burning). This value has been proposed as a definition for the boundary between brown dwarfs and giant planets. S Ori 47 could well be the result of a natural extension of the process of cloud fragmentation down to the deuterium-burning mass limit; a less likely alternative is that it has originated from a protoplanetary disk around a more massive cluster member and was later ejected from its orbit because of interacting effects within this rather sparse (~12 objects pc^-3) young cluster.^83a Some of these observable effects should include:

S Ori 47 also displays atmospheric features that is indicative of low gravity - weak alkaline lines with hydride and oxide bands - which are consistent with the expectation for a very young object still undergoing gravitational collapse. LiI (Lithium Iodide) is present in its atmosphere, which confirms the youth and substellarity of S Ori 47. Near-infrared photometry (J band) and low-resolution optical spectroscopy (600-1000 nm) conclude that S Ori 47 is a true substellar member of the s Orionis cluster. It has been proposed that the IR excess and high Li abundance observed in 4-8 % of the G and K giants originate from the accretion of a giant planet, a brown dwarf or a very low-mass star.^83b

APMPM J0559-2903, the coolest extreme subdwarf known, is unlike other small metallic dwarfs in that there are no very late type objects known among the extreme subdwarfs. Measured to be esdM7, APMPM J0559-2903 was discovered in a new southern high proper motion survey (Keck Telescope) The NextGen grid of model atmospheres by Hauschildt et al. (1999) determined that the effective temperature was ~3100 K and the metallicity was MH=-1.5. The theoretical parameters place this object at a distance of 100 pc with a space velocity of 260 km s(-1) relative to the local standard of rest.^83g

Astronomers have photometrically monitored (Cousins I_C) eight low-mass stars and brown dwarfs that are probable members of the Pleiades. The derived rotation periods for two of the stars - HHJ 409 and CFHT-PL 8 are found to be 0.258 and 0.401 days, respectively. The masses of these stars are near 0.4 and 0.08 M_solar, respectively; and the latter is only the second such object near the hydrogen-burning boundary for which a rotation period has been measured. The relative amplitude in the two bands shows that the spots in that star are about 200 K cooler than the stellar effective temperature of 3560 K and have a filling factor on the order of 13%. With one possible exception, the remaining stars in the sample do not show photometric variations larger than the mean error of measurement. HHJ 409 and the V_Mag 9.5 disk star 2MASS J0149090+295613 were also examined in the Visible spectral range. These had previously exhibited a strong flare event profiles, but at the time of this study, no photometric variation was detected.^83d

Hubble Space Telescope has made near-infrared NICMOS observations of a remarkable low-luminosity Class I (protostellar) source in the Taurus Molecular Cloud. IRAS 04325+2402 exhibits a complex bipolar scattered light nebula. The central continuum source may be multiple, or it may be crossed by a small dust lane. Complex arcs seen in scattered light surround the central source. While the physical nature of these structures is not clear, they may reflect perturbations from multiple stellar sources or from a time-dependent mass ejection. A second source is found at a projected distance of ~ 1150 AU from the central region, near the edge of a nebular lobe probably produced by outflow. This second source is another low-luminosity young stellar object, seen nearly edge-on through a dusty disk and envelope system with disk diameter of about 60 AU.

Theorist believe that the scattered light ``streaks'' associated with this second source are limb-brightened outflow cavities in the dusty envelope, possibly perturbed by interaction with the outflow lobes of the main source. The nature of the companion is uncertain, since it is observed mostly in scattered light, but is thought to be a very low mass star or brown dwarf, with a minimum luminosity of approximately 10^-2 L_solar. These protostellar (accreting brown dwarf ?) sources may have multiple centers of infall, nonaligned disks, and outflows, even on relatively small scales.^83e

Observations of the center of the Galactic globular cluster NGC 6273 produced a BV color-magnitude diagram (CMD) for ~28,000 stars. The most prominent feature of the CMD is the extended horizontal-branch blue tail (EBT) with a clear double-peaked distribution and a significant gap. The EBT of NGC 6273 was compared with the EBTs of seven other globular clusters for which we have a CMD in the same photometric system. The results of this comparison show that all the globular clusters with an EBT show at least one gap along the horizontal branch, which could have similar origins.

A comparison with theoretical models suggests that at least some of these gaps may be occurring at a particular value of the stellar mass, common to a number of different clusters. From the CMD of NGC 6273 a distance modulus (m-M)_V=16.27+/-0.20 was derived. An average reddening E(B-V)=0.47+/-0.03 shows the CMD is strongly affected by differential reddening, with the relative reddening spanning an E(B-V)~0.2 mag.^83f

M dwarf systems often have multiple components which include spectroscopic as well as wide binaries. Since brighter substellar companions are expected in younger systems, they should be associated with Me (hydrogen emission) dwarfs and flare stars which display coronal activity. Self-luminous planets and brown dwarfs are expected to show the best contrast against the sky background and the scattered light from their primary stars near their thermal emission peak.

Expected luminosities make young giant planets favorable for the first direct detection of an extrasolar planet. The giant planet formation process is relatively slow with expected formation times ranging from comparable to the star formation timescale up to the nebula lifetime, depending on the formation theory. Therefore quantitative models of giant planet formation have to be considered when estimating young giant planet properties at pre-main sequence stellar ages. This is especially important for free-floating giant planet candidates where planetary nature is inferred from luminosities, without independent mass determinations. To infer planetary nature in a young population, the properties of young planets have to be compared to those of proto brown dwarfs, as they follow from the respective formation theories.^83h The faint near-infrared low luminosity flare stars around Me dwarfs at high galactic latitudes were obtained at J, K' and a narrowband filter (K") to isolate the strong methane feature between 2.3-2.4 microns. ^83k

One of the more significant results from observational astronomy over the past few years has been the detection of low-mass companions (LMCs) to solar-like stars. The commonly held interpretation of these is that the majority are ``extrasolar planets" whereas the rest are brown dwarfs, the distinction made on the basis of apparent discontinuity in the distribution of M_sini for LMCs as revealed by a histogram. Results from formal statistical analysis of M_sini, as well as the orbital elements data for available LMCs test the assertion that the LMCs population is heterogeneous. The outcome was mixed. Solely on the basis of the distribution of M_sini a heterogeneous model is preferable, although no unique best-fit mixture can be determined. On the basis of the distribution of orbital periods and eccentricities a homogeneous model is strongly preferable. Overall, a definitive statement asserting that LMCs population is heterogeneous is perhaps unjustified. A remarkable statistical similarity was noted between compatible sample of stellar binaries and LMCs populations. This similarity coupled with marked populational dissimilarity between LMCs and acknowledged planets would suggest a common origin for all LMCs and selected stellar binaries as an alternative to the prevailing interpretation.^83l

A recent model for the stellar initial mass function (IMF), in which the stellar masses are randomly sampled down to the thermal Jeans mass from hierarchically structured prestellar clouds, predicts that regions of ultracold CO gas, such as those recently found in nearby galaxies by Allen and collaborators, should make an abundance of brown dwarfs with relatively few normal stars.

This result comes from the low value of the thermal Jeans mass, which scales as M_J~T²/P^1/2 for temperature T and pressure P, considering that the hierarchical cloud model always gives the Salpeter IMF slope above this lower mass limit. The ultracold CO clouds in the inner disk of M31 have T~3^o K and pressures that are probably 10 times higher than in the solar neighborhood. This gives a mass at the peak of the IMF equal to 0.01 M_solar, well below the brown dwarf limit of 0.08 M_solar. Using a functional approximation to the IMF given by [1-e^(-M/M_J⁾^²]M^-1.35d_logM for M>M_J, which fits the local IMF for the expected value of M_J~0.3 M_solar, an IMF with M_J=0.01 M_solar in M31 has 50% of the mass and 90% of the objects below the brown dwarf limit.

The brightest of the brown dwarfs in M31 should have an apparent extinction-corrected K-band magnitude of ~30 mag in their pre-main-sequence phase. For typical star formation efficiencies of <=10%, brown dwarfs and any associated stars up to ~2.5 M_solar should not heat the gas noticeably, but if the IMF continues up to arbitrarily high masses, then the star formation efficiency must be <=10^-4 to avoid heating from massive stars.⁸³ⁿ

Observations have already identified more than 25 sources with temperatures cooler than the latest M dwarfs. A comparison with detailed model predictions (Burrows & Sharp 1999) indicates that these L dwarfs have effective temperatures between ~2000+/-100 K and 1500+/-100 K, while the available trigonometric parallax data place their luminosities at between 10^-3.5 and 10. Those properties, together with the detection of lithium in one-third of the objects, are consistent with the majority having substellar masses.

The mass function cannot be derived directly, since only near-infrared photometry and spectral types are available for most sources, but VLM/brown dwarf models can be integrated into simulations of the solar neighborhood population and constrain Psi(M) by comparing the predicted L dwarf surface densities and temperature distributions against observations from the Deep Near-Infrared Survey (DENIS) and 2 Micron All-Sky Survey (2MASS) surveys. The data, although sparse, can be represented by a power-law mass function, Psi(M)~M^-a, with 1<a<2. Current results favor a value nearer the lower limit. If a=1.3, then the local space density of 0.075>M/M_solar>0.01 brown dwarfs is 0.10 systems pc^-3. In that case, brown dwarfs are twice as common as main-sequence stars but contribute no more than ~15% of the total mass of the disk.^83u

The discovery of four field methane ("T''-type) brown dwarfs using Two Micron All-Sky Survey (2MASS) data has provided new and useful information. One additional methane dwarf, previously discovered by the Sloan Digital Sky Survey, was also identified. Near-infrared spectra clearly show the 1.6 and 2.2 mm CH₄ absorption bands characteristic of objects with T_eff<~1300 K as well as broadened H₂O bands at 1.4 and 1.9 mm. Comparing the spectra of these objects with that of Gl 229B, thus, it is believed that all new 2MASS T dwarfs are warmer than 950 K, in order from warmest to coolest: 2MASS J1217-03, 2MASS J1225-27, 2MASS J1047+21, and 2MASS J1237+65. Based on this preliminary sample, a warm T dwarf surface density of 0.0022 T dwarfs deg^-2, or ~90 warm T dwarfs over the whole sky detectable to J<16. The resulting space density upper limit, 0.01 T dwarfs pc^-3, is comparable to that of the first L dwarf sample from Kirkpatrick et al.^83o

There is also the discovery of an isolated brown dwarf with similar properties to the binary object Gliese 229B and to the newly discovered field brown dwarfs from the SDSS and 2MASS surveys. Although exhibiting similar colors, its magnitude of ~ 20.5 is about 6 magnitudes fainter than Gliese 229B. This is the most distant of the several methane brown dwarfs reported to date, at a distance of ~ 90 pc. Its IR spectrum, although at low S/N given the faintness of the object, is remarkably similar to those of the other methane brown dwarfs.^83p The mass-luminosity relation from assumed mass functions and the luminosity functions of Jahreiss & Wielen and Gould, Bahcall & Flynncan be confirmed by comparison of the resulting mass-luminosity relations with data for binary stars constrains the permissible mass functions.

The presence of CH₄ in the atmosphere of Gl 229B and the near absence of CO indicate that its effective temperature is less than ~ 1000K. Atmosphere models reveal that at such a low temperatures, nitrogen should be in the form of N₂ and NH₃, with the latter favored at lower temperatures and at higher pressures. While N₂ is spectroscopically invisible, NH₃ has several strong absorption bands, most notably near 2mm and in the 10mm region.

A high-resolution, high signal-to-noise K band spectrum of Gl 229B which was fit with synthetic spectra to constrain the effective temperature, the surface gravity, the metallicity and the NH₃ abundance of Gl 229B was studied, and it was determined that the atmosphere of Gl 229B is believed to have a subsolar H₂O abundance. NH₃ is detected in the spectrum but with a NH₃/H₂O ratio of about 25% of the value expected from chemical equilibrium. This suggests that the NH₃ to N₂ chemical equilibrium is quenched at a non-equilibrium value by dynamical processes in the atmosphere.⁸³ⁱ Gliese 229B exemplifies the characteristics native to the giant planets. Optical and near-IR spectra reveal a reduced chemistry largely in thermochemical equilibrium. Disequilibrium processes such as condensation and vertical mixing are also suggested by the CO feature and the continuum of Gl229B's spectrum.

However, Gl229B's spectra show that brown dwarfs are also exotic, having properties unique to their genre. The temperature of Gl229B's upper atmosphere is too high for the condensation of ices seen in planets, and too low for the condensation of metals and silicates of stars. The resulting clarity of this atmosphere affords a view to exceptionally high pressures. Gl229B displays features of an element (Cs) present only at temperatures of brown dwarf atmospheres, which is sensitive to the equilibrium chemistry, the heavy element abundance, and condensation processes. In addition, the atmospheric structure appears special: Gl229B's high gravity causes the stratosphere to extend down to a pressure of ~ 20 bars, in contrast to the ~ 0.5 bar radiative-convective boundary observed for planets.^83j

If the mass function is assumed to be a power law, the best-fitting slope lies either side of the critical slope, alpha=-2, below which the mass in low-mass objects is divergent, depending on the luminosity function adopted. If these power-law mass functions are truncated at 0.001M_solar, the contribution to the local density from stars lies between 0.013 and 0.10M_solarpc^-3 depending on the mass at which the mass function is normalized and the adopted value of alpha. Recent dynamical estimates of the local mass density rule out stellar mass densities above ~0.05M_solarpc^-3. Hence, power laws steeper than alpha=-2 that extend down to 0.001M_solar are allowed only if one adopts an implausible normalization of the mass function. If the mass function is generalized from a power law to a low-order polynomial in log(M), the mass in stars with M<0.1M_solar is either negligible or strongly divergent, depending on the order of the polynomial adopted.^83q

Astronomers working on another effort to catalog the heavens have reported that they have imaged the coolest body ever found outside our solar system, a type of brown dwarf that could be as abundant as any other stars. The temperature of the object, which is larger than a typical planet but smaller than a star, is 900 Fahrenheit (500 Celsius). That makes it the coolest star-like object ever imaged beyond our solar system. Gliese 570D is a Jupiter-sized brown dwarf that lies just 18 light years from Earth as part of a previously known triple-star system in the constellation Libra. Gliese 570D was the 12th such object found by the Two-Micron All Sky Survey (2MASS) since 1995.

Astronomers have tentatively called the new objects methane brown dwarfs because of the presence of the gas in their spectra. Methane is typical of giant planets like Jupiter, but never (or very rarely) appears in stars or in most other brown dwarfs, both of which are far too hot. All told, astronomers worldwide have found some 15 of the methane brown, or "T," dwarfs, most in the last six months.

David Golimowski, a co-discoverer of the first methane brown dwarf, Gliese 229B, said the cool, dim objects could be as plentiful as the more brilliant stars that stud the skies in our stellar neighborhood. "There could be one for every star," said Golimowski, an associate research scientist at Johns Hopkins University.^53f Astronomers could reasonably expect to find even cooler brown dwarfs in the future, although present search methods are geared toward finding hotter, brighter objects.

Gliese 570D was discovered in a photograph made in 1998 as part of 2MASS's efforts to image the entire sky at near-infrared wavelengths. This discovery was later confirmed to be a brown dwarf by emission-line spectroscopy using the four-meter telescope at the Cerro Tololo Interamerican Observatory in Chile.

A team of European astronomers has found a cold and extremely faint stellar object in interstellar space. Located high above the galactic plane, it is a Methane Brown Dwarf, of which only a few are known to exist. It is also the most distant Methane Brown Dwarf identified to date. Methane Brown Dwarfs are the coolest members of the Brown Dwarf class detected so far, with temperatures ~ (approximately) 700 °C/1292 ^oF- which is ~1000 degrees cooler than the coldest stars. The new object, NTTDF J1205-0744, was found during a deep survey of a small sky region in the constellation Virgo just south of the celestial equator where chances of identifying a rare object like this in such a restricted area are very small.

The La Silla ESO 3.58-m New Technology Telescope (NTT) made a long series of exposures of a small sky field in Virgo between 1997 and 1998. Their goals was measuring and demonstrating the limiting performance of two astronomical instruments at this telescope, the SUperb-Seeing Imager (SUSI) in the visible part of the spectrum (0.35 - 1.00 µm), and the multi-mode Son of ISAAC (SOFI) in the near-infrared region (1.0 - 2.5 µm).

The NTT Deep Field - an area of observed sky that measures only 2.3 x 2.3 arcmin², has been studied in great detail during the first period of VLT observations. Such distant objects are quite red (due to their high redshift) and are best detected by a combination of visible and infrared exposures.

ESO PR Photo 35a/99		Caption to ESO PR Photo 35a/99: Part of the NTT Deep Field, with the new Methane Brown Dwarf NTTDF J1205-0744 at the centre. The field measures 1.3 x 1.3 arcmin². The object is well visible in the SOFI infrared exposure (left) in the J-band at wavelength 1.25 µm, but not in the SUSI one at a shorter wavelength (right) in the i-band at 0.8 µm. North is up and East is left.
The astronomers (at La Silla) noted a star-like object of extreme colour in this field. While it was well visible and similarly bright in both SOFI infrared images (J = 20.2 and K = 20.3), it could not be seen at all on the SUSI images in the visible spectral region, even at the longest wavelength (i-band) observed with that instrument (i-J > 6 mag), cf. PR Photo 35a/99. No "normal" object is known to have such extreme colours. The new object now received the designation NTTDF J1205-0744, indicating that it was discovered in the NTT Deep Field at the given position on the sky. It seemed that there were only two possibilities. Either it was an extremely distant quasar (redshift about 8) at the edge of the observable universe, or it must be a very cold object in the Milky Way Galaxy. Whatever its nature, this was obviously a most interesting object.		[Preview - JPEG: 400 x 251 pix - 72k] [Normal - JPEG: 800 x 502 pix - 224k] [High-Res - JPEG: 3000 x 1881 pix - 1.7M]

ESO PR Photo 35b/99

Caption to ESO PR Photo 35b/99: The infrared spectrum of NTTDF J1205-0744, as obtained with SOFI at the NTT and ISAAC at VLT ANTU, and compared to the spectrum of the much closer and brighter Methane Brown Dwarf Gliese 229B.
- This issue was resolved by obtaining infrared spectra of NTTDF J1205-0744. Despite its faintness, initial observations with SOFI at the NTT covering the infrared J and H-bands already revealed some of the molecular absorptions characteristic of methane brown dwarfs.
- More recently, complementary longer wavelength observations with ISAAC at the first VLT 8.2-m Unit Telescope (ANTU) at Paranal have now confirmed the nature of this object. The combined SOFI/ISAAC infrared spectrum shown in PR Photo 35b/99 is clearly extremely similar to that of Gliese 229B, the first Methane Brown Dwarf discovered a few years ago and which is a member of a binary system at a distance of about 19 light-years.
- The features in the spectra result from strong absorption by methane (CH₄) and water (H₂O). There is thus no doubt that NTTDF J1205-0744 is of the same type (stellar class T). Unlike Gliese 229B, however, it does not appear to be a member of a binary system. It is also 5-6 magnitudes (i.e., a factor of about 250) fainter than this and a few similar objects discovered recently in large-area sky surveys, implying that it is considerably more distant.

Properties of NTTDF J1205-0744

NTTDF J1205-0744 is located at a distance of about 300 light-years (90 pc) and some 240 light-years (75 pc) above the plane of our Milky Way galaxy. Its mass is probably about 20-50 times that of Jupiter, or less than 2% of that of the Sun. Its temperature is around 700 °C (1000 K), suggesting an age of 500 to 1,000 million years. Lacking a stable source of energy at its centre, it is becoming continuously fainter and cooler and will continue to do so for tens of thousands of millions of years.
- NTTDF J1205-0744 is a very faint and small object indeed, on the still not well understood border zone between stars and planets [1].

[Preview - JPEG: 400 x 337 pix - 56k]
[Normal - JPEG: 800 x 674 pix - 124k]

How many Brown Dwarfs?

How many T-class objects are there in the Milky Way? What is the space density of these extreme objects? Since only a few have been identified so far, any statistics must be quite uncertain. Until now, the best estimates have been of the order of 1 per 3,500 cubic light-years (0.01/pc³).

A surprising aspect of this discovery is that NTTDF J1205-0744 was found within a sky area of only 2.3 x 2.3 arcmin², specially selected to be as "empty" as possible in order to facilitate studies of distant galaxies. Based on the above density estimate, the chance of finding such an object should only have been about 1%. Based on model predictions, the chance would have been even smaller than this.

Searches like the one described here, based on the combination of optical and infrared data, therefore appear particularly effective at detecting such objects. It is now of high interest to test if this first discovery was just extremely lucky, or if the space density of these extreme objects is in fact much higher than expected.

Up until this time, the occurrence of planetary systems around brown dwarfs has not been studied in detail. However there are not reasons to assume that brown dwarfs have not planetary companions. Moreover, their smaller mass with respect normal stars makes more prominent the effects due the presence of the planets, allowing in principle the identifications of planets with smaller mass (i.e. Earth-like planets). A feasibility study of detection of planetary companions around the brown dwarf GL 229B (M=0.03 - 0.55 M_solar) using near-infrared high resolution spectroscopy to measure radial velocity variations has determined that the observational distinction between planets and brown dwarf secondaries is unclear.^83r Lacking observations relating directly to formation (e.g., multi-component system structure) the best one can do is to compare the overall properties of these sub-stellar mass companions (SSMC's) to those of the relevant stellar population: stellar-mass secondaries in binary systems (for which a nearly complete and unbiased survey exists).

Other than mass, all measurable properties of SSMC's to solar-like primaries are statistically indistinguishable from those of secondary stars in binary systems of similar primary spectral type (mid-F to mid-K dwarfs). In both cases semi-major axes, periods, and angular momenta are distributed approximately as f(x)? x^{-1} for x = a,P,L; and eccentricities approximately as f(e)? e^{-0.5}. There were no correlation of eccentricity with other orbital properties, aside from circularization of close orbits in both populations, detected. Secondary masses in both populations are uncorrelated with all orbital properties, the mass spectrum of SSMC's is seamless, and shows no statistically significant evidence for bi-modality reflective of a mixture of planets and brown dwarfs, which argues for a common formation mechanism for stellar-, and nearly all sub-stellar-, mass secondaries to solar-like stars - including objects of apparently planetary mass. In this sense the recently discovered "extrasolar planets" more resemble brown dwarfs than they do planets.^83s

A CCD-based photometric survey covering 870 arcmin² of a young stellar cluster around the young multiple star S Orionis. The limiting R, I, and Z magnitudes are 23.2, 21.8, and 21.0, respectively - the completeness being about 2.2 mag brighter. From color-magnitude diagrams 49 faint objects (I=15-21 mag) were selected, which smoothly extrapolate the photometric sequence defined by more massive known members. Adopting the currently accepted age interval of 2-10 Myr for the Orion 1b association, in which S Orionis is located, and considering recent evolutionary models, our objects may span a mass range from 0.1 down to 0.02 M_solar, well within the substellar regime. Follow-up low-resolution optical spectroscopy (635-920 nm) for eight candidates in the magnitude range I=16-19.5 shows that they have spectral types M6-M8.5, consistent with what is expected for true members.

Compared with their Pleiades counterparts of similar types, H_a emission is generally stronger, while Na I and K I absorption lines appear weaker, as expected for lower surface gravities and younger ages. In addition, TiO and in particular VO bands appear to be clearly enhanced in S Ori 45 (M8.5, I=19.5), compared to objects of similar types in older clusters and the field. The estimated mass of this candidate is only 0.020-0.040 M_solar hence, it is one of the least massive brown dwarfs yet discovered. The potential role of deuterium as a tracer of both substellar nature and age in very young clusters can provide a determination of the age of a cluster. The S Orionis cluster is an excellent site for determining this transition zone empirically. The most massive brown dwarfs identified are expected to have burned their deuterium content, while the lowest mass ones should have preserved it.^83t