This is a summary of a discussion on the science cases for a very wide-field IR imager for the Mayall 4-m, held on 5 February 1999. The following attendees participated in the science discussions:
Jeff Valenti
Steve Strom
James Rhoads
Joan Najita
Mike Merrill
Ken Hinkle
Dick Joyce
Arjun Dey

Finding Methane Brown Dwarfs with NEWFIRM

Science Questions:

• What is the IMF below the brown dwarf limit?
• Do brown dwarfs contribute significantly to local dark matter?
• Are isolated brown dwarfs more common than companion brown dwarfs?
• Can we find brown dwarfs in clusters, the Galactic disk, and the halo?
• What is the kinematic behavior of field brown dwarfs?
• What is the metalicity distribution of nearby brown dwarfs?
• What is the age distribution of nearby brown dwarfs?
• Which models reproduce observed broadband IR surface fluxes?
• Do brown dwarfs show anamolous abundance patterns?
• What are the chemical pathways of dust formation in brown dwarfs?
• What is the effect of dust on the structure of brown dwarfs?
• Why do some brown dwarfs show signs of magnetic activity?

Spectroscopic Overview:

Oppenheimer et al. (1998, ApJ, 502, 932) present the best available 1-2.5 micron spectra of M, L, and brown dwarfs. vB10 (Teff=2500 K), GD 165 B (Teff=1850 K), and Gl229B (Teff=900 K) demonstrate how the spectrum changes with descreasing temperature. Water bands increase in strength, perhaps becoming photometrically detectable, despite telluric absorption. For temperatures below about 1200 K, methane bandheads at 1.6 and 2.2 microns create distinctive features detectable with 5% wide filters at 1.50-1.58, 1.62-1.70, 2.05-2.15, and 2.20-2.30 microns. In Gl229B, the flux contrast across the bandheads is a factor of 4 (1.5 mag) at 1.6 and 2.2 microns, and a factor of 16 (3 magnitudes) at the long wavelength end of the L band.

Burrows et al. (1997, ApJ, 491, 856) present low resolution synthetic spectra of brown dwarfs as a function of mass and age. The range of molecular line depths present in the synthetic spectra generally agrees with the observed range of line depths described above. Assuming the models are correct, one can use the appearance of the spectrum to constrain mass or age. If a distance can be measured, then the surface flux provides another constraint, but it is not clear to me that this additional observable breaks the degeneracy between age and mass. The effect of abundance variations has not been explored in any detail. The synthetic spectra are used to construct color vs. absolute magnitude diagrams. The J-K vs. M(K) diagram is of particular interest because brown dwarfs are quite red (J-K=1.8 for GD165B) until methane absorption actually makes J-K blue (J-K=-0.2 for Gl229B). Custom narrow band filters would presumably provide even better selection criterea for very cool brown dwarfs. Finally, tabulated absolute magnitudes are useful for estimating sample volumes for limiting magnitudes in standard infrared filters.

Expected Detection Rates:

Buell Jannuzi argued in his "Near and Far in the Near-IR" proposal that if the undetected mass in the solar neighborhood (0.05-0.1 Mo/pc³) is all brown dwarfs (M/Mo=0.05), then the space density of brown dwarfs could be as high as 1/pc³.

Gl229B was the first brown dwarf shown to exhibit methane absorption. Several more have now been found by 2MASS and SDSS. Burgasser et al. (1998, AAS, 193.9803) discuss the approach by the 2MASS team to discover more methane brown dwarfs. In 1500 square degrees, 22 stationary sources were found with J-Ks<1.0, R-Ks>7, J<16, and Ks<15. Only 6 of the 22 sources had J-Ks<0.8, and none had J-Ks<0 like Gl229B. Spectroscopic followup is underway. Burrows (1997) claims that for a 900 K brown dwarf, M(J)=M(K)=14 for log(g)=3.5 (10 Myr) and M(J)=M(K)=16 for log(g)=5.0 (3 Gyr). Thus, the 2MASS experiment described above was sensitive to young brown dwarfs in 4% of the volume out to 16 pc and old brown dwarfs in 4% of the volume out to 6 pc, finding 0 to 6 brown dwarfs with strong methane absorption. At best, the space density of *methane* brown dwarfs in the solar neighborhood is roughly 0.2/pc³.

Several warmer brown dwarfs have been found in nearby young clusters. Most of the groups have used R and I bands to identify candidates, but some work has been done in the IR. Zapatero, Martin, & Rebolo (1997, A&A, 323, 105) give the following absolute magnitudes for objects at the substellar limit in nearby clusters and the field:

Specific Questions:

1. What are the field of view and spatial pixel scale requirements for this science project?

Large field of view obviously improves detection efficiency. The only constraint on pixel scale is the need for good isolated point source photometry in the H and K bands.

2. What is the depth / sensitivity requirement?

Based on the 2MASS analysis described above, we expect no more than 0.004/deg² methane brown dwarfs brighter than K=15. A detection rate of 1/deg² would require sensitivity down to H=19 and/or K=19.

3. What is the wavelength coverage requirement? ( > 2.5 microns?)

Viable methane bands exist in the H, K, and L bands. The bands are redundant, so one could sacrifice wavelengths beyond 2.5 microns and still find methane brown dwarfs. Followup spectroscopy at all wavelengths will be important, but that does not bear on the design of NEWFIRM.

4. What is the spectral resolution requirement?

R=20 (5% filters) should be adequate for identifying methane brown dwarf candidates and estimating effective temperature. Confirmation and detailed analysis of these rare objects can be done using followup spectroscopy.

5. Does this project require IR observations? (i.e., can it be done more efficiently at other wavelengths?)

Methane brown dwarfs have R-Ks > 7. Optical observations would have to be roughly 100 times more sensitive (or cover that much more sky per exposure) to be competitive. Moreover, the optical spectrum of Gl229B is relatively featureless, perhaps due to dust in the atmosphere, making optical selection criterea less reliable than direct IR measurement of methane bandheads.

6. Would the project benefit if the camera provided narrow-band filters or some slit/slitless IR spectroscopic capabilities?

5% custom filters are adequate for finding methane brown dwarfs and estimating temperatures. Candidate selection should be robust, so followup spectroscopy will probably be more efficient than trying to obtain slitless spectra of all sources in the field.

7. Does the immediate scientific success of this project require support observations at other wavelengths? (Ground-based / space-based observations?)

Simply generating a high probability candidate list has scientific value, especially once the contamination rate has been established. Followup spectroscopy in the IR would be the next logical step and should be quite feasible if the contamination rate is low. Optical spectroscopy will be useful for studying the properties of dust.

8. How does this project feed into / depend upon future observations at larger telescopes?

Large telescopes will be required to obtain followup spectra at the maximum possible spectral resolution.

The YSO Case for NEWFIRM

Key Problems:

• Quantify the star-forming history of nearby molecular
  • star-formation rate vs. time
  • IMF vs. time
• Quantify the range of lifetimes for the disk accretion phase
• Catalog candidate of post-accretion phase YSOs as a first step toward examining disk properties (evolution of solid and gas phase) during the period when solar systems develop and mature

Current Status:

• "shallow" optical proper motion surveys are available for
  • parts of the Taurus clouds (limit: M > 0.15-0.2 Msun)
  • parts of the Orion association (variable limits, but typically well above the hydrogen-burning limits)
  • NGC 2264 (M > 0.2 Msun; t < 3 Myr)
• these surveys do NOT provide:
  • a census of the obscured/embedded population associated with the cloud
  • a census of young stars NOT projected onto the main body of the star-forming region as traced by the molecular cloud
• accretion disk candidates can be selected from uv-excess; IR-excess; and H emission, though the depths of these surveys is variable
• x-ray surveys select (a) young (t < 3 Myr) disked and non-disked stars; (b) interloper field stars that are x-ray bright but could be old (t >= 100 Myr). Hence, though helpful, they are neither complete nor definitive in selecting the young stellar populations associated with molecular clouds.

Primary Survey Measurements:

• JHK fluxes, positions and proper motions for all objects K < 18; J, H < 20 in northern hemisphere molecular clouds with d <= 500pc
  • photometric precision of 0.03 mag/filter will enable (a) detection of sources with M > 0.08 for Av < 10 and t < 10 Myr; and (b) initial sorting of objects into disked/non-disked candidates from location in the JH/HK diagram
  • proper motions with a single epoch precision of 5 mas will enable robust sorting among objects born in molecular clouds and those foreground and background to these complexes on baselines of 3 years (for star-forming complexes located within 150 pc) and 10 years for the more distant regions.
• NOTES: improvements in single epoch precision may be possible
• required aerial coverage: ~ 100 square degrees

Potential Importance of the Survey:

• with proper motions, this survey will for the first time enable isolation of candidate objects associated with star-forming events spanning the anticipated lifetimes of molecular cloud complexes (10 - 20 Myr). In turn, this catalog of candidate members will provide the primary database for understanding the star-forming history and the early evolution of circumstellar disks.
• the catalog of accurate positions, proper motions and fluxes will provide an essential tool for selecting guide/acquisition stars and off-axis WFS stars to enable precise pointing and AO- correction for studies of selected objects in molecular cloud regions

Ancillary Observations:

• complementary optical surveys to (R,I) ~ 25 would enable more robust selection of accretion disk candidates via their location in the RI/IK color-color diagram.
• followup multi-object NIR spectroscopic surveys would enable spectral classification, which when combined with JHK photometry would yield Av, L, T (and thus masses and stellar ages) as well as a direct measure of K-band excess -- thus providing the basis for accurately locating stars in the HRD and developing a more complete list of disk candidates
  NOTE: spectroscopic followup is essential in order to address the key problems outlined above.
  the total number of targets for which spectra will be required number between 3000 and 10000.
• L-band surveys of selected regions would enable evaluation of the completeness of K- band searches for disk candidates (the combination of inclination effects and putative "inner disk holes" may render K-band searches incomplete)

Comments re FOV:

• typical cloud sizes range from several to several tens of square degrees
• the depth of the survey (K < 19) does not preclude using a "moderate" aperture telescope and a "moderate" field of view
• the requirement for proper motions probably sets a lower limit to field size based on the projected surface density of stars K < 19 seen THROUGH optically opaque regions (Av > 30 mag) regions of these clouds. I haven't checked to see whether we might be driven to FOV > 10 arcmin by the requirement to have sufficient stars within a field to "tie into" adjacent, less obscured fields.

Extragalactic Science for NEWFIRM

1. Evolution of Clusters of Galaxies

Relevant questions:

• Do high-redshift (z > 1) clusters exist?
• What is the evolutionary history of clustering? (i.e., is there a z dependence on the cluster-cluster correlation function?)
• What is the evolutionary history of the galaxy LF in clusters? (this provides a discriminant between hierarchical and monolithic galaxy formation scenarios)
• What are the properties of galaxies in clusters? (this is a large telescope project to study e.g. the fundamental plane in clusters as a function of z)

Proposed projects:

• survey for clusters
  • Near-IR search techniques for z > 1 clusters complement X-ray searching techniques. IR surveys are superior to optical surveys because of the confusion with foreground / background sources. Also, IR surveys are tuned to find the more evolved (redder) populations. Finally, the K-band flux from a galaxy is a fairly good measure of its mass even at high redshift.
  • Surface densities of high-z clusters are pretty unknown. Old estimate by Evrard & Henry 1991 for clusters with X-ray fluxes > 10^-14(in the Rosat band) and z > 1 : 5.3 to 18 clusters per steradian
  • Better estimate for clusters at z < 1 from Deeprange ~ 25 clusters / deg² If we are dealing with pure counting statistics, then for a simple statistic like the cluster-cluster correlation function, you need 50 - 100 clusters / bin for 10-15% error (purely from the Poisson stats)
  • ==>need several square degs (at least 10)
• The search phase of this project may be independently carried out using X-ray satellites, but that does not address the issue as to the relation / evolution of the X-ray gas compared with that of the cluster galaxies.

Specific answers to template questions:

• FOV requirements: ~0.5 deg
• need to cover >~ 10 deg²
• pixel scale: ~0.3 - 0.5arcsec/pix
• need good star/galaxy separation
• depth requirement: K~22 (==> need 4m) L* + (1 to 2 mag)
• spectral resolution: broad band filters (JHK)
• need IR? yes (contrast, mass measurements, LF)
• NB filters? not necessary
• Spectroscopic capability? Yes, for follow-up / confirmation
• Need other wavelengths right now? Not essential, but optical complement would be nice for cluster id & phot-z
• future: clean samples for deep spectroscopy and morphological evolutionary studies with Gemini, NGST

2. Formation and evolution of the red-envelope galaxies

Relevant questions:

• When and how do Es form (hierarchically or monolithic collapse)?
• What are the ages of high-z ellipticals?
• What is the galaxy LF in the field at high-z and how does it evolve with z?
• EROs: What are they? What is their space density? How are they relevant to galaxy formation / evolution? Are they related to E galaxies (progenitors?)?

Discussion:

The formation of the oldest galaxies provide a direct lower limit to the age of the Universe and thereby independent constraints on the cosmological parameters
Main problem: we don't know of many high-z `ellipticals' (galaxies dominated by old populations) primarily because these tend to be missed in optical surveys. An L* elliptical at z=2 has K=21, R-K=8 ==> IR surveys are the best way to pick out these objects.
Also, there exists a red population which is only just being uncovered. Very little is known about these objects. Surface density estimates vary from 0.01 - 1 per arcmin².
Areal coverage required: unknown, but probably large. For objects with I-K > 4 & K < 20, surf.dens. = 0.26 per arcmin² (Barger et al.) = 2.5 per arcmin² (EES) For objects with R-K > 6 & K < 17.5 surf.dens. = 0.01 per arcmin² (Hu & Ridgway 1994) = 0.1 per arcmin² (Beckwith et al 1998) For objects with R-K > 6 & Ks < 20 surf.dens. = 0.7 per arcmin² (EES)
May imply strong field-to-field variations / clustering (Pretty uncertain)
Note that if we want to work down the LF, we really need to get to K=21 or 22.
J-K is a pretty effective way of selecting out the old high-z populations, and is also a very rough z discriminant (for z < 4, J-K is roughly linear with z for red-dead pops).

Proposed projects:

• survey for red-envelope galaxies 3-color survey complemented with one optical band can pick out the red envelope objects and separate the Es from the EROs (EROs tend to have redder colors). This could well be the same survey as that for clusters above.
• follow-up red/IR spectroscopy with large telescopes to determine zs, ages (for ellipticals), etc.
• follow-up with space-based opt/IR telescopes for morphology, and IR/sub-mm telescopes for LW SED (EROs are supposed to be a major component of the sub-mm bkgd)

Specific answers to template questions:

• FOV requirements: need to cover >~ 1 deg²
• pixel scale: ~0.3 - 0.5arcsec/pix need good star/galaxy separation
• depth requirement: K~22 (==> need 4m) L* + (1 to 2 mag)
• spectral resolution: broad band filters (JHK)
• need IR? yes (contrast, mass measurements, LF)
• NB filters? not necessary
• Spectroscopic capability? Yes, for follow-up / confirmation (need large telescope for the Es. EROs could be done on 4m with efficient IR spectrograph to detect H Need other wavelengths right now? at least one optical band (although J-K or I-K could suffice)
• future: clean samples for deep spectroscopy and morphological evolutionary studies with Gemini, NGST, SIRTF, WIRE

3. Primeval galaxy searches

Relevant questions:

• What is the earliest epoch of galaxy formation / existence?
• What is the space density of high-z emission line galaxies?
• How dusty are highz star-forming galaxies?

Proposed project:

• survey in narrow-bands + broad-bands for line emission + follow-up spectroscopy (or narrow-band opt observations for z-exclusion (in [OII], say))
• This is a high risk, but high payoff project

Specific answers to template questions:

• FOV requirements: unknown! expect need to cover >~ 1 deg² (could estimate from theory) Previous surveys by Bunker covered 30 arcmin² to 2 X 10^-16 erg/s-cm² and found 1 object
• pixel scale: 0.5arcsec/pix
• flux sensitivity requirement : 10^-17 erg/s-cm² (L=10⁴⁴ erg/s at z=13)
• spectral resolution: narrow band filters (0.1 - 1%)
• need IR? yes (redshift)
• NB filters? essential
• Spectroscopic capability? Yes, for follow-up / confirmation (may need large telescope for most of the fainter lines) Possible to do brighter ones on 4m with efficient moderate resolution IR spectrograph)
• Need other wavelengths right now? at least one optical narrow band for redshift discrimination
• future: clean samples for deep spectroscopy and morphological evolutionary studies with Gemini, NGST

4. Complementarity to ongoing / planned NASA projects:

• object identification and classification
• astrometry with high positional accuracy (esp for IR satellites)
• follow-up for variable / moving targets

Bottom line requirements:

• FOV: ~0.5 deg FOV to efficiently survey fields ~ 1-10 deg²
• pixel scale : 0.3 - 0.5 arcsec/pix for star/gal separation
• sensitivity: K~22, J~24 for broad-band filters ~ 10^-17 erg/s-cm² for narrow-band filters
• requires the aperture of the 4-m
• no need for contiguous FOV as long as this can be constructed efficiently