Dear Readers,
Welcome to issue 8 (December 14, 1998) of the MIPS/IRAC GTO newsletter. In this issue, Paul Richards provides summaries of deep galaxy surveys with ISO and Keck. Karl Stapelfeldt and Mike Werner present a summary of a half-day nearby star workshop held at the end of the recent SWG meeting in Pasadena. And William Reach presents some numbers on the zodiacal background, which will be important for planning deep surveys with SIRTF.
The deadline for submissions to the next issue of the newsletter is Sunday the 3rd of January, 1999. Have a merry Christmas and a happy new year.
Doug Kelly, editor (dkelly@as.arizona.edu)
The following is a summary of a discussion with Marc Davis regarding the DEEP project. DEEP is a dense redshift survey of galaxies with redshifts between 0.7 and 1.5, using the Deimos spectrograph on Keck II with a 15 arcmin FOV. The spectral coverage is 350 nm with resolving power of 3700. Four regions will be surveyed, each 120x15 arcmin, beginning in summer '99 and taking 2-3 years. Redshifts of 30,000 galaxies are expected. This survey is described on the website http://www.ucolick.org/~deep/
DEEP is conceived as part of a massive assault on the universe at redshift unity. Science goals include: spectral properties of galaxies versus luminosity and redshift; measurement of 2- and 3-point correlation functions; coherence of large scale structure at z=1; measurements of small scale thermal motions. A number of follow-up observations are planned including:
It would be best for our team to become well informed on this project. The issue of field selection for low dust regions could be especially important. The DEEP fields are in principle chosen and are listed below. They are taken from a preprint by Marc Davis and Sandy Faber, which is available at http://xxx.lanl.gov/abs/astro-ph/9810489
RA dec (epoch 2000) 14h 17 +52 deg 30 Groth Survey Strip 16h 52 +34 deg 55 last zone on low extinction 23h 30 +0 deg 00 on deep SDSS strip 02h 30 +0 deg 00 on deep SDSS strip
The following is a summary of the ISOCAM deep surveys, which are a precursor to the proposed MIPS deep survey. Information was obtained from conversations with J.L. Puget 11/9/98 and a presentation by Bill Reach at the SWG on 11/13/98. Relevant papers are Guiderdoni et al. 1997 (Nature, 390, 257), ISO Deep Far-Infrared Survey in the Lockman Hole, Kawara et al. 1998, A&A, 336, L9, and FIRBACK: A deep survey at 175 microns with ISO, preliminary results, Puget et al. (preprint).
ISOCAM 6.5 and 15 microns pin a minimum in the cosmic infrared background (CIRB). More sources are seen at 15 than at 6.5 microns. 30% of sources are like M-51; only 5% are like Arp 220. For wavelengths beyond 250 microns, all workers agree on the CIRB. At 100 microns there is some disagreement due to the method of correction for the ionized component.
Longward of 6.5 microns, the integrated background is 31 nW m-2 sr-1; shortward, it is 17. Compare to Soifer and Neugebauer (1991, AJ 101, 354), who found that the infrared luminosity in the local universe at wavelengths between 8 and 40 microns is ~40% of that at wavelengths > 40 microns.
Kawara et al. used ISOPHOT at 95 and 175 microns to survey two 40' square fields in the Lockman hole. Their results simulated the FIRBACK survey. The proposal is described in Guiderdoni et al. The overall survey was 120 hours at 175 microns over 4 square degrees.
Marano Fields (X-ray, Radio & ISOCAM)
Saturn scans were used to measure the point spread function. There was a big contribution from the wings, which increased throughput by a factor of 2.2. With a correction for the PSF, the Marano field observations give a background level within 13% of the DIRBE result.
A preprint has been prepared by Puget et al. on the Marano 1 field; other fields will be published later. They covered an area of 0.25 sq degree to a detection level of ~100 mJy (5 sigma) and produced a catalog of 23 sources. This count total is many more than predicted from the IRAS luminosity function without strong evolution. Sky confusion is the dominant source of noise. Number counts are a factor 2-3 above models with evolution in Guiderdoni et al. Sources above 120 mJy contribute 0.1 MJy/sr, which is 10% of the submillimeter background. Extrapolation of counts to 8.5 mJy provides the whole background.
The power spectra predictions by Guiderdoni et al. must be corrected. The predicted cirrus curve is about right, the detector noise was 10 times less than predicted, and the power spectrum due to sources is 10 times higher than predicted.
Puget feels strongly that the longest possible wavelength filter should be used. Their c-160 filter gave 175 microns on galaxies and 160 microns on dust.
The Lockman hole project was started too late to be repeated, so Puget et al. observed the ELAIS fields in the south at 175 microns.
No results are available yet from these fields.
Here is a summary report on the nearby star workshop held on Friday 13 November in conjunction with the SIRTF SWG meeting. Please feel free to contact either of us if you have questions or comments. The top level summary is that the discussions were very productive and have pointed the way to a number of activities that can be productively coordinated among the SIRTF instrument teams and the General GTOs. The focus of the workshop was towards debris disks and main sequence stars rather than towards protoplanetary disks and pre-main sequence stars.
Karl Stapelfeldt krs@wfpc2.jpl.nasa.gov
Chas Beichman chas@ipac.caltech.edu
Dale Cruikshank dale@ssa1.arc.nasa.gov
Erick Young eyoung@as.arizona.edu
Mike Jura jura@clotho.astro.ucla.edu
Bob Gehrz gehrz@ast1.spa.umn.edu
George Rieke grieke@as.arizona.edu
John Stauffer stauffer@cfa.harvard.edu
Lee Hartmann hartmann@cfa.harvard.edu
Bill Latter latter@ipac.caltech.edu
Tom Roellig troellig@mail.arc.nasa.gov
Jeff van Cleve vancleve@astrosun.tn.cornell.edu
Michael Werner mww@ipac.caltech.edu
On very short notice, George Rieke provided a summary of the discussions. He broke down SIRTF's science goals for nearby stars into three major categories and listed specific project observational goals in each. They were:
There are several specific issues that need to be addressed in the next few months as we move forward. We would like to request that we work along the following lines and distribute draft material or write-ups by March 1, in time for an iteration prior to the April SWG meeting. The first listed person is responsible for making this happen (the others listed are being volunteered by us):
(1) Develop version zero target lists of candidate nearby stars for study. (Werner, assisted by Stapelfeldt and Van Cleve). Discussion is needed of our selection criteria and detection strategies. See what the upcoming NStars database will provide and what additional research we need to do. This would include both a list of prominent disk systems for detailed study ["The Dirty Dozen" - a phrase coined by Chas which was adopted instantly by all present] as well as a set of targets for a more general survey. It should include zero order definition of observing modes and estimates of integration times.
(2) Develop toy models of debris disks to calculate synthetic observations, especially images. (Stapelfeldt, assisted by Jura and Hartmann). Produce a tool that can build our intuition about how to observe these systems.
(3) IRAC team should give their assessment of phase space for brown dwarf detection vs. groundbased near-IR surveys (Stauffer - or Stauffer designee in view of impending launch of SWAS - calling on Beichman for assistance regarding L dwarfs as needed).
The first speaker was Mike Werner, who stated the purpose of the meeting was to initiate a coordinated approach to SIRTF GTO science observations of the circumstellar material and substellar companions of nearby stars. "Identify the most promising science programs, people who are interested in pursuing them, and preparatory work which needs to be done" Nine speakers made the following presentations:
Mike W. then began an overview of the MIPS debris disk program, which he has worked out in conjunction with Chas Beichman. This broke down into three specific proposals:
1. The "dirty dozen" most prominent nearby debris disks (Vega, Fomalhaut, Epsilon Eridani, Beta Pic, etc.) should be singled out for extensive study: MIPS photometry, MIPS superresolution imaging; IRS at both low and high resolution; and IRAC imaging for substellar companions. Mike pointed out the value of these observations as a "ground truth" foundation for subsequent debris disk work and as early and exciting demonstrations of SIRTF's scientific capabilities to the public. George Rieke asked what criteria would be used to define the sample size and to select the as-yet unspecified sources 5-12 in this program.
2. The Nearest Stars. These offer the best opportunity to spatially resolve disks and companions from the central star. They could have disks which can't be discerned photometrically against the photosphere but which could be detectable in images as extended structures in the PSF wings. These sources represent our best chance to detect very tenuous debris structures, perhaps down toward the limit of 1 solar zodi (tau = 10^-7). Mike W. showed a simulation of such a detection; Chas expressed some doubts that saturation effects could be overcome in making these observations. Likely observations would be super-resolution imaging with MIPS, IRS in low and high resolution, and IRAC imaging for companions. The possibility of detecting disks in scattered light with IRAC was discussed, and there was a general feeling that this would not be easy to do as well as it could be done from the ground with a large telescope at K.
3. The Nearest Stars of Various Types. This program (a.k.a. "The statistics of Dust Disks" would address the question of how disk and companion properties might be related to those of the primary star. Aiming for a sensitivity to circumstellar dust of tau= 10^-6 (10 zodies in a Kuiper debris disk), this program would examine groups of stars in each spectral class A-M, groups with binary stellar and substellar companions (the latter sources being those with planets detected in radial velocity surveys), and a group selected according to various indications of youth. This program might include hundreds of stars selected for high galactic latitude in order to minimize background cirrus confusion. Possible observations were MIPS photometry and SED mode, to be followed up by IRS and IRAC in appropriate cases. Mike W. presented rough estimates of the needed exposure times; reaching the tau=10^-6 limit is clearly going to be expensive for M stars. In light of this, and given the uncertainties in the photospheric emission from M stars, George asked if there was a clear need to do M stars at all. Mike Jura pointed out that Poynting-Robertson drag would be much less effective at removing particles from debris disks around M stars, and this might lead to interesting differences between these and the debris disks already studied around earlier type stars. Karl Stapelfeldt muttered that the frequency of debris disks around M stars is very poorly constrained from the IRAS data and is something we'd like to learn from SIRTF.
In the summary discussion, George noted that some study is needed to identify the best wavelengths for imaging these disks. The disks appear larger at longer wavelengths, but the stellar PSF does also. He suggested that we might not know until IOC how PSF stability, detector saturation, and ghosting would affect these observations, and thus the best strategy might be to take some early pilot observations and assess how best to proceed from there.
John reviewed the expected progress to come from WIRE AI programs which will provide a better census of substellar objects in nearby molecular clouds. He said that IRAC will continue this work to study isolated brown dwarfs in young clusters. He will also be pursuing this through an aggressive ground-based photometry program of open clusters. For nearby stars, the IRAC GTO program will include observations of about 50 targets to search for companions. Most of these will be M stars, as these will minimize the contrast challenge for detecting faint companions. They are also considering work on a second group of perhaps 50 objects within 100 pc, and selected to be especially young (ages of a few hundred Myrs). This second group is currently being explored in the near-IR by the Kulkarni group using Keck. A third group of objects that would be attractive to study is the L stars found by the 2MASS survey, to see if they have companions. (An L star as a binary companion has already been detected by Rebolo et al. 1998). And finally, the systems in which planets are known from radial velocity measurements are obvious places to look for additional substellar companions. This discussion led to the question posed above about the uniqueness of SIRTF's capabilities for detecting brown dwarf companions to identified stars [again in comparison to what could be done from a large ground-based telescope]. Stauffer also pointed out that many of the nearby stars that might be included in the programs described above will have been thoroughly searched for companions [via both direct imaging and radial velocity measurements] by the time SIRTF launches.
Chas gave an overview of 2MASS progress; the survey is about 40% complete in so far as acquiring data is concerned. The "rare object" project has invalidated its own definition by finding a lot of field brown dwarfs: 20 now seen in only 0.9% of the sky. Objects are flagged as interesting if they have unusual R-K colors. Every candidate with colors like GL 229B has turned out to be a spurious moving object, and thus no methane BD has yet been seen by 2MASS. Chas estimated that only a few dozen of the CH4 dwarfs may turn up in the whole 2MASS survey. The objects they are seeing are all L dwarfs. He then described the spectral characteristics of these new objects, showing that they lose the TiO and VO bands prominent in M stars but add new features such as metal hydrides. An L0 star corresponds to Teff of 1900 K, and an L8 star to Teff of 1200 K. None of these contain CH4 bands, which must turn-on very quickly (perhaps over just a delta T = 100 K) at the L9 class and beyond. Chas referred to GL 229B as a "T dwarf" object, a designation which has not yet been adopted by the IAU.
Chas said that trig parallaxes are beginning to become available for a few of the L dwarfs, and the distances are turning out to be 10-30 pc. The ages are not known for any of the objects specifically, and thus it is not really possible to derive individual object masses from theory. Instead mass functions are discussed to reproduce the range of luminosities seen. Despite the uncertainties, it's clear that the total mass of L dwarfs falls well short of solving the cosmological missing mass problem. Nevertheless, indications are that there could be more L dwarfs in the galaxy than all other types of stars. He mentioned the possibility that the Sun's nearest neighbor might not be Alpha Cen but rather an as-yet undiscovered L dwarf. He suggested that SIRTF should aim to look for L dwarfs as companions to nearby stars, and let WIRE (or SIRTF serendipity) find the field objects.
There was a short discussion about a database for the properties of nearby stars. Dale Cruikshank commented that the Hipparcos catalog now has the most accurate distances to nearby stars, but that it does not go as faint as the Gliese catalog. Some merger of the two catalogs would be needed. Karl mentioned a recent email exchange with Dana Backman, who said that the NASA funded NStars database will become available in initial (very preliminary) form in January 1999.
Dale outlined the current state of knowledge about the Kuiper Belt: 65 known objects. Their orbits are roughly known, and a broad range of inclinations is seen. The orbital eccentricities are quite pooly known, however, since the objects have not been observed over enough of an orbit arc yet to derive this quantity. Their sizes range from 75-590 km, if the assumed albedos of 5% are correct. There is a claim of millions of 10 km (comet nucleus sized) objects in the Kuiper Belt based on HST detections of an excess of noise spikes moving in a prograde direction (Cochran et al.), but this is very controversial. There are two dynamical classes of objects: the Plutinos, which are nearer objects like Pluto trapped in a 3:2 resonance with Neptune; and more distant objects whose orbits are uncorrelated with Neptune's. Dale made the interesting remark that the term "Plutino" is a linguistic nightmare, being "the latinized diminuitive of an anglicized greek word"
The key question Dale raised was whether knowledge of our Kuiper Belt would aid understanding of debris disks, or whether the information transfer would work in the opposite direction. He mentioned that the dust component of our Kuiper Belt has not been directly measured due to the difficulty in isolating it from the bright foreground zodi emission at r= 2-3 AU. By analogy, SIRTF observations of other G stars might shed light on this question. Mike W. said that this was a good justification in pursuing measurement to the one-zodi sensitivity limit (tau= 10^-7) in as many stars as possible. Note that the Kuiper Belt is the structure in our Solar System most analogous to the debris disks seen around stars like Vega and HR4796.
The discussion then turned to how SIRTF measurements could advance understanding of our own Kuiper Belt. The spectra of Kuiper belt objects are difficult to obtain even with Keck, but indicate a diversity of properties as seen in reflected light. The presence of strong spectral features in some objects seems to conflict with a trend among asteroids for high albedo objects to show strong spectral features. This calls into question the albedos being assumed for KB objects. SIRTF can measure the thermal emission from KB objects, and from radiative balance arguments the albedo can then be derived as was done by IRAS in the case of main-belt asteroids. Dale presented some sample calculations in his SWG viewgraphs that showed that useful photometry can be done with MIPS in about 500 sec on a typical KB object. Spectroscopy, however, could be quite expensive in terms of integration time, but is important if we are to improve our knowledge of the composition of the Kuiper Belt. George said that it would be profitable to study spectra of the captured satellites of the outer planets (Neptune's Nereid, Saturn's Phoebe, etc.) as brighter analogs to KB objects. Dale agreed, and he pointed out that the KB objects scattered into nearby but unstable heliocentric orbits (Chiron and Pholus) could also serve this purpose. Dale argued especially for observations of Pholus, as its lack of cometary activity suggested a pristine surface.
John gave what he said is the "pessimistic view" of how well we can know the ages of field stars. The standard indicators used for age include the strength of chromospheric emission from Ca II lines; the strength of X ray emission; the stellar rotation period or vsini; the photospheric lithium abundance; a star's location on the HR diagram relative to theoretical isochrones; space motion association with other stars whose age is known; and metallicity. None of these methods is as good as we would like; several only work for later type stars. Some only work in certain narrow ranges: example, X ray luminosity of a star saturates above a particular vsini (or stellar age). He said that any claims to know the ages of field stars to better than a factor of 2 are overly optimistic.
All the age dating techniques discussed above are tied to the ages of open clusters derived from isochrone fitting, and their ages are only known to a factor of two. He pointed out that a new age for the Pleiades (125 Myrs) has recently been derived using the properties of recently discovered brown dwarf members. This is significantly older than the previously accepted value of 70 Myrs and may point to improper treatment of convective overshoot in previous pre-main sequence star interior models. He also made the point that HR diagram isochrones are essentially uncalibrated for ages younger than the Pleiades, since there have not been enough dynamical measurements of pre-main sequence binaries to give the theory a sound foundation.
The age of HR 4796 was discussed, which John says is established by the presence of lithium in the low-mass companion. Models predict that Li in M stars will be completely burned in about 10 Myrs, which thus sets the age of that system at about 8 Myrs. The Li abundance vs. age calibration is different for every spectral type, since the depth of convective mixing of photospheric material is the key variable.
Mike J. told us that all stars tend to be a little bit lower in their observed mid and far-IR flux densities than might be expected from a simple Rayleigh-Jeans law. He said that we are unlikely to know the photospheric brightness to better than about 10%, and thus the smallest believable IR excesses measured with SIRTF will be at about 20% of the photospheric level. He did a sample calculation showing this limit for an F star, with the result being that the "floor" for a Kuiper debris disk detection lies at about 10^-5 Earth masses of dust (or approximately 10 zodies).
A short discussion then took place about SIRTF standard stars. George said MIPS will probably be calibrated using K stars, and 5% photometric accuracy is the instrument team's goal. Mike J. warned that IRAS K giants appeared to vary in the far-IR at the 10% level; that cirrus confusion would have to be properly accounted for; and that any ISM local to the K giants might be warmed by the stars, producing excesses not related to the photosphere OR bound circumstellar material. Karl worried about modest IR excesses due to debris disks around the calibration stars, and whether SIRTF could avoid experiencing the calibration experience IRAS had with Vega.
George pointed out that uncertainties in the stellar spectra could be normalized out using standards with the same spectral type as our science targets. Weak silicate or other features might reveal themselves in a ratio of target and standard star spectra taken with IRS.
Jeff highlighted the IRS capability to identify optically thin emission from olivine grains in debris disks. Understanding the variation in grain mineralogy among different debris disks, and their relation to cometary grains in our solar system, will be a primary goal for IRS work on nearby stars. He also pointed out, and several in the group agreed, that the IRS peak-up camera offers an opportunity to make 15 micron images with a pixel size of 1.8". Although intended primarily for target acquisition, the peak-up array offers unique and important SIRTF imaging capability. We encourage the Science Center in its plan to enable its use.
Jeff mentioned that IRS is capable of observing K dwarf photospheres out to distances of 30 pc. The IRS GTO program will include a study of the spectra of late-type stars; this data should help to understand the photospheric emission (and thus how well we can discriminate real excesses) throughout SIRTF's wavelength range.
Karl compared the sensitivity of SIRTF/IRAC to existing ground-based near-IR searches for brown dwarf companions. He considered a fiducial object with mass 20 Mjupiter and age 1 Gyr, as a companion to a main sequence M5 star. He used the spectral models of Burrows et al. 1997 Ap.J. 491 856 to get the expected brown dwarf colors. The contrast of this CH4 BD against the star is 230 at 1.6 microns, about 100 at 4.5 microns (the preferred wavelength for IRAC work), and about a factor of 10 at the MIPS wavelengths. This is a more favorable case than GL 229B, where the primary is an M1 star and the contrast is a factor of 10^4 in the near-IR; for earlier type primary stars, the contrast requirements become still more severe. IRAC has the sensitivity to detect the fiducial brown dwarf in a few seconds of integration time, but IRAC's abilities to image a high-contrast scene are not yet quantified. By looking at IRAC's PSF (FWHM of about 2.5") Karl guessed that contrast required to detect the fiducial companion would only be achieved outside a radius of 2". In the near-IR, any moderate-sized (3 meters or larger) groundbased telescope can gather enough photons to detect the same fiducial brown dwarf in a minute or less of integration time. The CFHT Adaptive Optics system has already demonstrated near-IR detection of objects at contrasts of >200 at radii of 0.5", which is sufficient to detect the same object as SIRTF/IRAC but to a much smaller limiting radius. When AO systems are eventually fitted to 8 and 10 meter telescopes, they will presumably do even better. Thus Karl concluded that for the brown dwarf atmosphere models that he considered, SIRTF did not have distinct advantages over the existing and soon-to-be-realized near-IR AO surveys for brown dwarf companions to nearby stars. SIRTF does have an important role to play in filling out the SEDs of these objects at longer wavelengths, however.
A key question is how much we can rely on the current brown dwarf
spectral models to represent the detectability of these objects across
the 1-5 micron region. Earlier in the meeting, John expressed some
confidence in the model colors for brown dwarfs. The group agreed that
the presence of dust or other condensates high in these atmospheres would
act to strongly suppress the "non-thermal" near-IR emission and
would hand a strong detectability advantage back to SIRTF. Nobody could say
how likely this situation might be. However, Werner adds in proof the
following note, which is that energy conservation would suggest that the
most extreme case would be one in which the brown dwarf spectrum becomes
a black body at its effective temperature, which should allow a pretty
simple determination of the advantage.
To Top
I made a table showing the minimum and maximum zodiacal light brightness as a function of ecliptic latitude. This might be useful for selection of fields for the GTO deep surveys.
These values were calculated using the COBE/DIRBE zodiacal light model (Kelsall et al. 1998 ApJ, in press).
For each ecliptic latitude, the minimum and maximum brightness were determined for all ecliptic longitudes and dates that satisfy the solar elongation constraint of SIRTF (80-120 degrees).
This table only applies to a wavelength of 12 microns, which I recommend as a "nominal wavelength" for SIRTF, in particular for selection of fields to observe and assessment of the relative background from one latitude or date to another. I have taken the approximation that SIRTF is very close to the Earth.
To understand the reasons for the variation in brightness, here is a capsule explanation:
The min and max in this table are in MJy/sr, and the ecliptic latitude is in degrees.
eclip. zodiacal
lat. min
max
____ ____ ____
0 21.4 46.3
5 20.0 43.6
10 18.8 40.5
15 17.3 36.0
20 15.9 31.6
25 14.6 28.0
30 13.6 25.0
35 12.7 22.7
40 11.9 20.7
45 11.2 19.1
50 10.7 17.8
55 10.3 16.7
60 9.9 15.8
65 9.9 15.2
70 9.9 14.7
75 9.9 14.3
80 10.0 14.0
85 10.2 13.4
90 10.4 13.0
To Top