Dear Readers,
Welcome to the October 5 issue of the MIPS/IRAC GTO newsletter. This episode features two new white papers. John Stauffer discusses how we can expand our knowledge of the stellar mass function by using IRAC to detect brown dwarfs down to 0.01 Msun and extrasolar bodies down to Jupiter-like masses. Bill Latter and Joe Hora discuss a combined IRAC/MIPS program to study stellar ejecta, with an emphasis on planetary nebula and supernova remnants and their interactions with the interstellar medium.
If you have comments on any of the papers in this newsletter, you can send them to me and they will be included in the next newsletter. The deadline for submissions to the next newsletter is Friday, October 16.
Doug Kelly, editor (dkelly@as.arizona.edu)
Until about two years ago, brown dwarfs were only a hypothetical construct - no reason was known why they should not exist, but no actual members of the class had been identified with certainty. Things have changed greatly now. The 2MASS program has identified several dozen brown dwarf candidates (Kirkpatrick et al. 1998) - and "confirmed" at least 10 of them with intermediate resolution spectra obtained at the Keck telescopes (where confirmation means showing the star has an extremely cool effective temperature and has a strong lithium absorption feature present). Since the 2MASS program has only analysed data for a very small fraction of the sky, it is expected that a few thousand field brown dwarfs should eventually be identified in the 2MASS database. A similar number of high probability brown dwarf candidates have been identified as members of nearby, relatively young open clusters and star-forming regions - with "confirmation" that at least 10 of them are very likely brown dwarfs again coming via Keck spectroscopy (Stauffer, Schultz and Kirkpatrick 1998).
Interpretation of the field brown dwarfs is not straightforward because the age of the objects is poorly constrained, as is the metallicity and distance (though it should be possible to determine parallaxes for a sample of the 2MASS brown dwarf candidates eventually). Searches for brown dwarfs in open clusters are particularly useful because membership in the open cluster means that the distance, metallicity and age of any identified brown dwarfs should, in principle, be known quite well. In fact, an age derived from the location of the lithium depletion boundary (the bolometric magnitude below which it is found that very low mass (VLM) cluster members have preserved their cosmic lithium abundance because the star's central temperature is below that needed for lithium fusion) is now believed to provide the most accurate, least model dependent age for an open cluster compared to any other means (Bildsten et al. 1997; Ventura et al. 1998).
The Pleiades has been by far the most popular target for open cluster brown dwarf searches because it happens to have a good combination of properties - nearby (d = 125 pc), rich (n ~ 800 members), high galactic latitude (b = 24 degrees), relatively compact (most of known members within 2 degrees from the cluster center), and relatively young (age as determined from the lithium depletion boundary is 125 Myr). Deep optical imaging surveys have now covered about 3 square degrees of the cluster to R = 23, I = 22, corresponding approximately to a limiting mass for cluster members of about 0.04 solar masses (Bouvier et al. 1998). The cluster mass function appears to be slowly rising still in the mass range 0.3 to 0.05 solar masses according to both Bouvier et al. and Zapatero-Osorio et al. (1997). Surveys at I and K may be able to extend the mass range to about 0.03 solar mass prior to SIRTF launch, though it is unlikely that such a survey will be able to cover as large an area as has been done in the optical. To date, the largest K band survey of the Pleiades for brown dwarf purposes covers of order 1200 square arcminutes (Schultz 1998).
The Hyades is three times nearer than the Pleiades, but 5-8 times older. In some sense, the nearness of the Hyades is actually a detriment to brown dwarf searches because it results in the Hyades having a very large angular extent on the sky - of order 100 square degrees. With of order 600 known members, one expects only about 6 stellar members per square degree, suggesting that one would need to survey many square degrees in order to identify a significant number of substellar mass objects. As of early 1998, no good brown dwarf candidates have been identified in the Hyades, though the faintest suggested Hyades members with I ~ 17 probably have masses slightly below 0.1 solar mass. Depending on the wavelength region in use, the nearer distance but older age for the Hyades relative to the Pleiades approximately cancel, and surveys to similar limiting fluxes will reach similar limiting masses in the two clusters.
At younger ages, the situation is both murkier and more intriguing. Stauffer and collaborators have obtained deep, large area optical imaging surveys of the Alpha Persei (age ~ 60 Myr) and IC2391 (age ~ 30 Myr) open clusters and have identified a large number of brown dwarf candidates. We have observing time this fall (including at Keck) to determine the lithium depletion boundary in both clusters and thus to determine the cluster ages accurately. In the process, we should identify brown dwarfs down to of order 0.04 solar masses. At ages less than about 10 Myr, the lithium test is no longer useful because at those ages lithium is predicted to still be present throughout the VLM/BD mass range. Therefore, identifying substellar objects with certainty is difficult. To date, such identifications have been made based solely on effective temperature - an object is assumed to be substellar if it is cooler than the stellar/substellar boundary identified in the Pleiades (basically arguing that the tracks are expected to be nearly vertical in an HR diagram). About half a dozen probable sub-stellar objects have been identified on this basis so far in star-forming regions (Luhman et al. 1997, 1998; Wilking et al. 1998; Comeron et al. 1998).
The contribution that can be made with IRAC to these studies is: (a) IRAC's ability to conduct large-area surveys efficiently and (b) IRAC's longer wavelength coverage, allowing IRAC to identify cooler (lower mass) brown dwarfs and to use the methane band signature in the 3 to 5 micron region as a signpost of very cool photospheres. For example, IRAC could survey a 4 square degree region of the Pleiades to a depth sufficient to detect brown dwarfs down to 0.01 solar masses in only of order 30 hours of telescope time. A similarly deep survey in a nearby, 1 Myr old star-forming region would be sensitive to isolated low mass objects down to the mass of Jupiter, if they exist. These surveys will not only extend the lists of detected objects to much lower masses than currently available, but they should also provide a more secure measure of the mass function by providing a census of objects over a larger fraction of the cluster or star-forming region than has generally been possible from the ground. The IRAC team intends to propose deep imaging surveys of several nearby open clusters and star-forming regions, with the goal to sample a range of ages and cluster types. The number of clusters and the area surveyed will depend on the total amount of observing time available to this project.
The target regions for the brown dwarf program will overlap partially with the regions that would be observed for the circumstellar disk programs envisioned by IRAC and MIPS. Therefore, it will be useful to merge these programs or at least plan them with cognizance of the needs of the other program.
References:
Bildsten, L. et al. 1997, ApJ 482, 442.
Bouvier, J., Stauffer, J. et al. 1998, AA 336, 490.
Comeron, F. et al. 1998, AA 335, 522.
Kirkpatrick, J.D. et al. 1998, in Brown Dwarfs and Extrasolar
Planets (R. Rebolo, ed.).
Luhman, K., Liebert, J. and Rieke, G. 1997, ApJLett 489, L165.
Luhman, K., Rieke, G., Lada, C. and Lada, E. 1998, ApJ 508, xxx.
Schultz, G. 1998, thesis, UCLA.
Stauffer, J., Schultz, G. and Kirkpatrick, J.D. 1998, ApJLett. 499, L199.
Ventura, P. et al. 1998, AA 331, 1011.
Zapatero-Osorio et al. 1997, ApJLett. 491, L81.
In recent years the study of stellar ejecta released during final evolutionary stages has provided many new and unpredicted insights into stellar evolution. The reason for these major leaps forward have come from the advancement of near-IR imaging and spectroscopy, as well as space-based platforms (HST, ISO) providing new views of the stellar death process. In its ability to observe objects in the mid and far infrared on the large scale, SIRTF will provide yet another jump in knowledge about very poorly observed, but key aspects of stellar ejecta from many relatively nearby objects.
A significant amount of material released to the ISM by planetary nebulae (PN) and supernovae (SN) is in the form of silicate and carbonaceous dust. Tiny carbon grains, or polycyclic aromatic hydrocarbons, have strong emission features throughout much of the IRAC and IRS accessible spectral regions -- in addition to numerous atomic and molecular features from photon dominated and shock heated gas. Thermal emission from cold (15 - 40 K), larger carbonaceous and silicate grains will produce significant emission observable by MIPS. Details of these emission components are very poorly known. One large PN was observed by ISO (The Helix; Cox et al. 1998). It was found that the mid-IR emission is dominated by emission from molecular hydrogen in the ground rotational state. Other successful ISO observations (some carried out by Hora and co-I's) will help to guide our SIRTF observations. SIRTF provides our last chance to explore the far-IR emission from PN and supernova remnants (SNR).
We suggest the following as possible science programs for MIPS and IRAC, with some IRS follow-up.
This aspect studies the overall structure and emission components of PN and SNR. Except for a few very bright, "young" sources done from the ground, very little study has been made of the long wavelength IR properties of stellar ejection remnants. Mass lost on the asymptotic giant branch is observable as extended envelopes around PN. It is this mass loss that is the dominant source of processed material into the ISM. It is of considerable importance to know the make-up of that material. Similarly, SN are the primary source of highly processed material. Much of that material quickly goes into the prodigious quantities of dust formed in the expanding shell. In addition, there are many poorly understood aspects of the late stages of stellar evolution that result in a menagerie of shapes of PN and SNR. Examining in detail the dust component can add key insights to when and how that shaping started.
There is considerable evidence now for mass loss variations at periods different from any known stellar pulsation times. In addition, some PN and SNR show evidence that such variations started very early in the evolution of the object -- before PN formation and SN explosion. Examining the cold dust can show us the extent of these shells, providing timescales and morphologies. Interaction with the ISM, especially for SNR, can have important implications to the evolution of structure in the ISM. This aspect of stellar ejecta has not been examined in any detail to date. However, some very interesting progress has been made in the study of Wolf-Rayet shells using IRAS data (J. Nichols, private communication). It is evident that stellar wind interaction with the ISM is important to the formation of large scale structures and star formation. MIPS can provide a detailed view of such interactions.
Object selection will be made from the rather sizable list of potential regions for study. We will optimize the use of the IRAC and MIPS (24 and 70 um) FOVs through source selection by minimizing maps of blank sky (regions of about 4' to 30' in size for MIPS). One set of objects (about 4' or less in angular size) will not require significant mapping and can use MIPS Super Resolution to increase spatial information. Overheads due to Observatory motions will be kept to a minimum with these objects. A second sample of objects could be chosen such that mapping will be required -- these would be objects >30' in total extent (the minimum MIPS scan length).
Several strategies are possible for this project. Because the potential sample size is very large, the best scenario will be determined by the total amount of time available to this program. Examples include: concentrate on big, faint PN and SNR such as the Helix Nebula (25x25 arcmin) to examine detailed structure; concentrate on small, faint PN like the Ring Nebula (about 3 arcmin) to assess the properties of the extended gas and dust components; examine a large statistically significant sample of small, bright PN to look at the faint extended emission. The bright sample will have cores that will saturate in all IRAC and MIPS bands, even with the high dynamic range option for IRAC. For some objects, it is likely that the MIPS 24 um data should be taken in the raw data mode rather than sample up the ramp (raw mode is available only to the instrument teams) to obtain the maximum dynamic range in ground processing. The most likely scenario will be a combination of those given above.
The imaging time required will depend strongly on the individual object or region being observed. The Helix Nebula provides one example. From the data taken by ISO, we find that, to cover the 25x25 arcmin region with the high dynamic range option (to avoid saturation of the brightest parts) and 100 seconds for the longest frame time with 2 dither points per raster position, about 2 hours total time for IRAC will be required. IRAS and other data have shown that most PN peak in flux at about 30 microns and are typically 2 orders of magnitude brighter at 30 microns than at 8 microns (Pottasch et al. 1984). These data indicate that the objects will, in general, be about 2 times fainter at 70 um that at 24 um. The 160 um fluxes are unknown but might provide the most information on the early evolution of these objects and their interaction with the ISM. From this, we find that a 30x30 arcmin scan map at the slow rate will observe the Helix to about the same sensitivity with MIPS as IRAC. In general, objects that fit into a single IRAC or MIPS FOV can be observed in about 10 minutes with IRAC. MIPS time will be comparable with super resolution measurements at 70 microns requiring multiple points for some objects. As for the imaging, observing times for MIPS SED and IRS spectra will depend strongly on the objects being observed, ranging from a few minutes to about an hour.
As a point from which to begin discussions on strategies, we suggest a good sample for both MIPS and IRAC imaging will be 15 to 20 large field objects (PN and SNR) from 2 arcmin to >30 arcmin in size. A second sample of "compact" objects requiring short integrations would be made up of about 80 to 100 objects. The MIPS and IRAC samples do not necessarily have to be identical. Sources for MIPS SED and IRS spectra will be selected to be a subset of the imaging samples.
It should be noted that, if previous experience holds, the large field objects that might be observed with SIRTF as part of this proposed program will likely provide some dynamic and interesting images with strong appeal to the general public and others.
Suggested Co-I's from outside SIRTF:
Aditya Dayal (IPAC/JPL)