Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.eso.org/~fcomeron/yipmo_df.ps Äàòà èçìåíåíèÿ: Fri Dec 1 20:27:41 2000 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:53:30 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: photosphere

Deep in a star forming region with the VLT:
looking for sub­Jupiter mass objects
F. Comer'on
European Southern Observatory, Karl­Schwarzschild­Strasse 2, D­85748 Garching
bei M¨unchen, Germany
Abstract. A remarkable result of evolutionary models for isolated objects with
temperatures of ¸ 1; 000 K is the prediction of near­infrared fluxes orders of mag­
nitude larger than they would have if emitting like black bodies. Model predictions
of exotic near­infrared colors due to prominent molecular bands have been con­
firmed by very cool evolved objects discovered in a variety of surveys, implying
that such very cool objects are both detectable as well as relatively easy to identify.
While such cool objects in the field are relatively evolved, with masses in the range
of tens of a Jupiter mass, objects with similar temperatures in star forming regions
should have masses in the Jupiter­Saturn range and could be detectable in deep
imaging of nearby star forming complexes. Their identification would be essential
both to demonstrate their existence and to test the input physics and chemistry of
the models.
Here I report on a deep JHK survey carried out with ISAAC at the VLT of
a selected area in the Chamaeleon I complex. The vast majority of the objects
detected have the colors expected from background high­z galaxies, but one faint
object stands out because of its nearly zero H \Gamma Ks color, while being undetected
at Js . The (H \Gamma K) color suggest a temperature of order of 1,000 K. The blue
(H \Gamma Ks) color may be due methane and molecular hydrogen absorption in the Ks
band, while the non­detection at Js might indicate an abundant presence of dust
in the photosphere.
1 Introduction
The search for the least massive substellar objects has been so far in the do­
main of very wide field or all­sky imaging surveys, aiming at the detection of
freely­floating nearby objects, or of targeted observations searching for com­
panions to known objects. With few and rather serendipitous exceptions ([6]),
the coolest substellar objects known to date have been discovered in dedic­
ated searches falling in those categories, rather than in deep field searches. By
a fortunate chance, direct identification and follow­up spectroscopy of objects
as cool as 1,000 K and even less ([2]) is taking place at a time when precise
modeling of objects with masses intermediate between those of the lowest
mass stars and of giant planets ([9], [3], [4]) is producing far­reaching predic­
tions testable by observations. The intense interaction between theory and
observations over the last decade provides one of the most spectacular recent
examples of mutual feedback leading to an accelerate pace of developments
in a field of astrophysics.

2 F. Comer'on
The masses of the coolest objects directly detected either in the field or
as companions are typically in the range of a few tens of a Jupiter mass.
The recent discovery of very late­type objects in the young aggregate around
oe Ori ([10]) pushes this limit towards even smaller masses, just a few times
the Jupiter mass. This narrows down the tantalizing gap between the masses
of the lightest freely­floating, directly detectable objects, and those of the
giant planets in our own Solar System and in the rapidly growing number of
extrasolar planetary systems ([8]).
Deep fields are needed to close this gap. Here I report on a deep infrared
image of a field in the Chamaeleon I star forming region where the first object
with a mass of the order that of Jupiter or even less may have been found.
2 Model predictions
The emergent flux redistribution due to molecular opacities at temperatures
in the 1,000 K range, mainly H 2 and H 2 O, enhances the flux in the near­
infrared to values several orders of magnitude above those expected from a
blackbody at the same temperature and luminosity. At the same time, the
strong enhancement at J s and the depression of the K band due to broad
CH 4 and H 2 O absorption bands gives these objects a uniquely blue infrared
color. These predictions have been essentially confirmed by observations of
T dwarfs ([1]) in the field and as companions to low mass stars.
Table 1. Expected photometric properties for isolated planetary­mass objects at
the ages of star forming regions (from Burrows et al. 1997)
Mass (M fi ) Age (Myr) MH (H \Gamma K)
5 1 9.3 1.22
1 1 13.2 0.08
1 5 16.1 ­0.22
0.5 1 15.9 ­0.48
The small sensitivity of these features to surface gravity ([3]) suggests
that objects with T­dwarf­like spectra in star forming regions should be also
easy to detect and identify, but their masses should be much lower at any
given temperature. Table 1 gives representative results with giant­planet­like
masses at typical ages in star forming regions ([3]). They should be taken as
a qualitative approximation, as large offsets in both absolute magnitude MH
and color (H \Gamma K) may be expected due to incompleteness in model input
quantities, in particular CH 4 linelists. As an aside, it must be said that strong
disagreements exist among different groups on whether or not such objects

Looking for sub­Jupiter mass objects 3
should be called ''freely­floating planets'' (as their masses are in the giant
planet range) or ''low­mass brown dwarfs'' (as they are isolated and have
formation histories and environments different from those of actual planets).
To avoid entering such discussions, I am hereafter referring to these objects
as ''young, isolated planetary­mass objects'', or YIPMOs for short.
3 Searching for YIPMOs with the VLT in the
Chamaeleon I star forming region
The center of the Chamaeleon I star forming region is a promising ground
for the search of YIPMOs. It is only 160 pc from the Sun ([7]), and contains
a low­extinction strip abundantly populated by young low­mass stars and
brown dwarfs, most of which are about 2 Myr old ([5]). Deep near­infrared
imaging was carried out on 27 and 28 March 2000 using ISAAC at the VLT
of a field in Chamaeleon I flanked by some of the brown dwarfs reported
in [5], while being sufficiently far from bright stars to avoid scattered light.
Deep exposures were obtained totaling 7 h 9 in K s , 2 h 5 in H, and 2 h 7 in J s .
The FWHM of the combined images (0''81 at J s , 0''71 at H, and 0''66 at
K s ) is good taking into account that, due to the very Southern declination
(\Gamma77 ffi ) of Chamaeleon I, all the observations had to be done at airmass near
2. Five­oe detection limits were J s = 23:0, H = 22:4, K s = 22:5.
A color­magnitude diagram of all detected sources down to H = 22:5 is
shown in the left panel of Figure 1. The color distribution shows a sharp blue
edge near 0.3, with many of the objects near it probably being background
stars. Redder objects, most of which are clearly extended in our images, are
mainly background galaxies at varying redshifts. However, one object stands
out at H ' 22:0, H \Gamma K ' 0:0. Despite its large formal 1­oe error bar given by
the PSF­fitting photometry (with a constant PSF derived from the brightest
stars in the field), the fact that the aforementioned blue edge of the color
distribution extends to magnitudes fainter than H = 22 suggests that the
actual errors are not underestimated, and that the gap between the color of
this object and the rest of the objects in the field is real. Therefore, based on
its unusual blue color, this object may be considered as a candidate YIPMO
in Chamaeleon I. The left panel of Figure 1 shows an image centered on it
in the J s , H, and K s bands, where it can be seen that, while all the other
objects in the field become more prominent as one goes from the H to the
K s frame due to both their intrinsic colors and the somewhat greater depth
of the K s image, the appearance of the candidate YIPMO remains similar
between both frames.
The left panel of Figure 1 also shows evolutionary tracks corresponding
to different masses and ages, from [3], at the distance of Chamaeleon I. The
tracks plotted here correspond to ages between 1 and 10 Myr and masses
between 0.5 and 5 M Jup . The merging of isochrones in a fairly narrow band
implies a considerable degeneracy between age and mass. More important is

4 F. Comer'on
Js
Ks
H
5 M at 1 Myr
Jup
2 M Jup
0.5 M Jup
at 10 Myr
at 1 Myr
Fig. 1. Left: (H \Gamma K);H diagram showing the position of all the objects detected
in those two bands in the imaged ISAAC field. The candidate YIPMO is the point
at the bottom left with formal 1­oe error bars derived from PSF­fitting photometry
attached. Curves are theoretical loci of YIPMO models ([3]); solid curves are iso­
chrones, and dotted lines mark equal masses. Right: a close­up view of the candidate
YIPMO identified in the ISAAC field in the Js , H, Ks bands. The extract of the
image shown here measures 23'' \Theta 15'', and is centered on the object.
the fact that models at any age predict colors about 0.5 mag bluer than those
of the candidate YIPMO at the same magnitude. This may well be however
a shortcoming of models rather than a real conflict between observed and
predicted properties of the YIPMO candidate: in fact, an offset in color by
the same amount and in the same sense is found between model results and
observations of field T dwarfs, thus supporting this interpretation.
An important discrepancy between model predictions and observations of
T dwarfs on one side, and the candidate YIPMO on the other, is the very red
(J s \Gamma H) or (J s \Gamma K s ) implied by the marginal detection at J s . Models predict
however (J s \Gamma K s ) ' 0:0, which is excluded by the observations. A tentative
interpretation may be based in the presence of dust in the atmosphere of
the object, producing an overall reddening of the emergent spectrum ([4]).
Models are somewhat uncertain about the role of dust in such objects: at the
relevant temperatures discussed here, dust should condense and precipitate
under the photosphere, while convection stops somewhat below the Ü = 1
level and in principle should not be able to bring the dust grains back into
the atmosphere. Nevertheless, it has been suggested that mechanisms such
as overshooting or turbulent difusion may alter the situation preventing the
complete disappearance of dust grains from the atmosphere. While the blue

Looking for sub­Jupiter mass objects 5
(J s \Gamma K s ) colors of field T dwarfs suggest that dust is indeed largely depleted
from the atmosphere, it may be possible that this is not the case for objects
within the ranges of ages and surface gravities discussed here. Confirmation of
the YIPMO nature of the object discussed here would thus provide important
constraints on the modeling of cool atmospheres.
4 Future prospects
The spectral energy distribution of the Chamaeleon I YIPMO candidate in
the near­infrared is clearly different from that of any of the other objects
detected in the field (where over 400 sources are detected in at least one
band), and cannot be reproduced by the continuum spectrum of a star or of a
normal galaxy at any redshift. While the possibility of it being a real YIPMO
is very appealing in view of its location in the direction of a star forming
region, alternative possibilities cannot be discarded with the available data.
The only one that appears viable is that of an object having strong emission
redshifted into the H and K windows. This may be the case if z ' 2:4, with
Hff+[NII] appearing at K and Hfi + [OIII] at H; or of z ' 3:4, with Hfi +
[OIII] at K and [OII] at H. Interestingly, although spectroscopy of such a faint
YIPMO may be hard to obtain, the alternative possibility of an emission­line
object may be readily confirmed or rejected with a few hours of integration
with an 8­m telescope, thanks to the bulk of the emission being concentrated
at a single wavelength. Spectroscopic follow­up is thus the obvious next step
in trying to verify the nature of this intriguing object. If its membership in
the Chamaeleon I star forming region can be confirmed, the gap between
directly observed very low mass objects and objects with truly giant­planet
masses will most likely have been filled.
Acknowledgements: I am pleased to thank the organizers for having al­
located time to make this presentation an oral one. Valuable comments and
supplementary data from Adam Burrows gratefully acknowledged, as well as
remarks on this work by Isabelle Baraffe and G¨unther Wuchterl.
References
1. Burgasser, A.J., et al., 1999, ApJ, 522, L65.
2. Burgasser, A.J., et al., 2000, ApJ, 531, L57.
3. Burrows, A., et al., 1997, ApJ, 491, 856.
4. Chabrier, G., Baraffe, I., Allard, F., Hauschildt, P., 2000, ApJ, 528, 397.
5. Comer'on, F., Neuh¨auser, R., Kaas, A.A., 2000, A&A, 359, 269.
6. Cuby, J.G., et al., 1999, A&A, 349, L41.
7. Knude, J., HÜg, E., 1998, A&A, 338, 897.
8. Marcy, G.W., Cochran, W.D., Mayor, M., 2000, in ''Protostars and Planets IV'',
eds. V. Mannings, A.P. Boss, S.S. Russell, Univ. of Arizona Press.
9. Tsuji, T., Ohnaka, W., Aoki, W., Nakajima, T., 1996, A&A, 308, L29.
10. Zapatero­Osorio, M.R., et al., 2000, Science, 290, 103.