Principal Investigator: James Larkin (Assistant Professor, UCLA)

Co-Investigator: Tiffany Glassman (UCLA)

This proposal describes a five year plan to use advanced adaptive optics systems and infrared instrumentation to study early galaxy formation and evolution. The ultimate goal of the scientific program is to study the morphology, metallicity and star formation rates of faint galaxies as a function of age (redshift) in the early Universe, and thereby determine when and how galaxies such as own our Milky Way formed. These observations will rely primarily on natural guide star adaptive optics and on the high surface density of faint galaxies on the sky. In preliminary observations of the fields of fifteen bright stars, we have already identified over one hundred candidate galaxies within their isoplanatic patches. We believe that these observations will grow and involve a much larger number of center members as AO instrumentation improves.

First-year project plan: An intrinsic problem with the earliest form of the Keck Adaptive Optics system is its reliance on natural guide stars brighter than about 12 magnitude at R. This limitation makes most extragalactic targets unobservable because of their intrinsic faintness and the relative rarity of bright enough guide stars. The P.I. has developed an interesting strategy for circumventing this limitation for extremely faint galaxies. These observations rely on the high density of galaxies on the sky. Deep infrared surveys (e.g. Djorgovski et al., 1995) have shown that there are about 2x10^5 galaxies per square degree brighter than K=24 mag. To a limiting magnitude of K=20 mag this number is down by about a factor of ten to 2x10^4 per square degree. This still implies that within 20 arcseconds of ANY guide star there are on average 2 galaxies brighter than K=20 mag and 20 galaxies brighter than 24th magnitude. Recent redshift surveys (e.g. Cohen et al., 1996) have shown that the average redshift of field galaxies brighter than K=20 mag is greater than 0.5 and this should only rise at fainter magnitudes. So the strategy is to perform deep infrared imaging around bright (<12 mag) A- type stars. The goal of these observations is to provide the highest resolution infrared images ever obtained of faint field galaxies. As discussed in the introduction, these observations should provide great insight into the structure and evolution of young galaxies. The first light camera has sufficient sensitivity to image faint galaxies, but its main problem is actually the relatively small field of view. It is only 4.3 arcseconds on a side and so cannot be randomly placed with the expectation of finding a distant galaxy. An obvious solution is to use the Keck non-AO infrared camera to image the fields around bright stars and identify candidate galaxies, and then use the AO system to achieve much higher resolutions. A sample of bluish stars that will pass close to the zenith has been prepared for these early observations. We have already imaged fifteen of these stars down to a K band magnitude of 22 mag, and identified over one hundred candidate galaxies within their isoplanatic patches, some as bright as 16 magnitude at 2 microns. So the primary activity of the first year of the center, will be to use the first light camera to take the highest resolution images of faint field galaxies ever. This requires a couple of nights on the Keck Telescope and funding travel, data analysis software and graduate student salaries.

First-year milestones:
Obtain diffraction limited images of several candidate galaxies with KCAM and the Keck AO system. These would be broad band images capable of determining morphological classifications, bulge to disk ratios, and potentially (via IR colors) a crude estimate of the age and redshift.

Second through fifth-year plan:
Morphologies alone are not sufficient for determining the properties of a distant galaxy. It is also extremely important to determine the redshift of a faint galaxy and to analyze the spectral features to detect the presence of young stars, dust obscuration or an active galactic nucleus. But as galaxies become fainter and further away, optical spectroscopy becomes more difficult both because many galaxies become redder, but also because the traditional diagnostic spectral features are shifted to longer wavelengths. NIRSPEC is a new powerful infrared spectrograph due for commissioning on the Keck Telescopes in October, 1998. As part of the commissioning plan and continuing into the regular 1999+ observing semesters, the P.I. (a member of the NIRSPEC team) will begin taking R~2000 spectra of faint field galaxies. Target objects will principally be selected from the AO fields around bright stars. At this resolution, it can achieve a 10 sigma detection of the continuum in each pixel in 1 hour for a 19.5 magnitude object. Much greater sensitivities are expected for line dominated sources. This is well matched to the typical candidate galaxies for the AO system. It is anticipated that 4 to 5 galaxies can be observed each night in one band. In the few spectra of faint field galaxies that exist, most spectra resemble either quiescent galaxies or starbursts with very few active galaxies discovered. This makes the resolution of 2000 ideal, since it gives the best sensitivity to narrow line components while also offering good wavelength coverage. With the combination of morphology and redshift, some statistics can be built up on the typical structural composition of galaxies as a function of age. By further comparing the spectra with galaxy models such as from Bruzual and Charlot (1993) it is hoped that the history of each galaxy can be deduced. The models show significant sensitivity to the infrared colors and spectral slopes depending on the age of the stellar population and the amount of internal reddening. Optical spectra of K selected galaxies such as Cohen et al. (1996) have shown that emission lines are more prevalent in high redshift galaxies, and it is expected that this trend will continue as we push further back towards the first generations of stellar formation. By analyzing such spectral lines, an enormous amount of information can be gained about a galaxy. Hydrogen recombination lines such as H? are good diagnostics of UV photon production connected to young stars and potentially to the presence of an AGN. The ratios of other emission lines such as Oxygen and Sulfur lines can determine the metallicity of early galaxies, and the gas densities and temperatures. A high power (~30 Watt) laser beacon system has been delivered to the Keck Observatory and is planned to be integrated with the AO system in 2000. Based on the density of appropriate guide stars and the expected performance of the AO system with the Laser, this will give a full sky capability with only a modest loss in system performance (~75% of the natural guide star peak intensity). This integration date also corresponds with the currently scheduled delivery date of NIRC2, a facility class infrared camera being built at Caltech. Prof. Larkin is a Co- Investigator on this camera and will play an important role in the early testing of this instrument. The camera has three plate scales of 0.01, 0.02 and 0.04 arcseconds per pixel in order to Nyquist sample the diffraction limit of the Keck Telescope at 1.25, 1.65 and 2.2 microns. The camera has a state of the art 1024x1024 InSb infrared array, sensitive from 1 to 5 microns, and a spectroscopic capability with a resolution of about 5000. Coupled with the natural and laser guide star systems, this instrument will provide an unparalleled infrared capability. It is expected to be the dominant instrument used during the middle of the proposal period. With the much larger field of the NIRC 2 camera (up to 40''), it will be possible to image large numbers of distant galaxies, rather than one galaxy at a time with the AO first light camera. When coupled with the laser guide star system, faint galaxies can also be imaged away from bright stars leading to increased sensitivities and the ability to do parallel studies with the Hubble Space Telescope and other space based missions (SIRTF, WIRE etc...). With the increased sensitivity of the AO system, much deeper infrared galaxy counts can be performed. The simple act of counting galaxies as a function of brightness can put strong constraints on the formation period and evolution of galaxies, and also on cosmological parameters. The deepest optical images such as from the Hubble Deep Field (Williams, et al., 1996) have detected a flattening of galaxy counts at very faint levels that is more pronounced at shorter wavelengths (Pozzetti, et al., 1998). This is at least in part due to the "dropping out" of high redshift galaxies as their lyman continuum is blocked by intergalactic hydrogen. Such a flattening has not been detected in the deepest infrared images to date (Djorgovski et al., 1995) and the lyman continuum is not shifted past one micron until a redshift of ten of more; certainly larger than any true protogalaxy. So infrared counts become critical for searching for faint galaxies from high redshifts. The ultimate goal of these observations is to study the process of early galaxy formation. To do this, we need not only morphological information about a distant galaxy but also internal kinematics. The NIRC 2 camera described above can determine some spectroscopic information about these distant galaxies, but there is a much more efficient instrument on the drawing boards. Prof. Larkin is the P.I. on a design study sponsored by the Keck Science Steering Committee to design an imaging spectrograph with built in OH-suppression capabilities. The instrument concept uses a bundle of 2048 optical fibers to sample a rectangular patch of the sky and produce a spectrum for each location. By combining the spectra, an image with any desired range of wavelengths can be produced and analysed, or spectral variations can be examined across the image. By dispersing at a resolution of 5000, the instrument separates the OH emission lines of the Earth's atmosphere that produce the dominant source of infrared background from 1 to 2.2 microns. By adding together only those spectral channels free from OH atmospheric lines, the resulting image or spectral cube has a much lower background and is expected to achieve the faintest imaging capability of any infrared instrument until the development of the Next Generation Space Telescope. The P.I. is developing this instrument specifically to target high redshift galaxies. By achieving high spectral and spatial resolution simultaneously, the instrument can examine the internal dynamics of very faint galaxies (K~22 mag). This capability is crucial in determining the role of mergers in early galaxy formation and evolution. As stated above, many faint galaxies (presumably at high redshifts) have highly distorted morphologies, but without kinematic information, it is impossible to determine if these are simply star forming regions orbiting a fairly stable galaxy or the remnants of merging galaxies.

Schedule of later milestones:

2000-2002 Use the first facility AO camera (NIRC2) to image larger numbers of faint galaxies and perform determine faint number counts and clustering properties. NIRC2 can also take AO assisted spectra of fainter galaxies than with NIRSPEC.

2002-2004. Integral field studies of evolving galaxies to study kinematic and morphological structures. This will produce the deepest infrared images prior to NGST.