Science Goals for Antarctic Infrared Telescopes
Michael G. Burton, John W.V. Storey, Michael C.B. Ashley, PASA, 18 (2), in press.

Next Section: Complementarity with other Facilities
Title/Abstract Page: Science Goals for Antarctic
Previous Section: Results from Site Testing
Contents Page: Volume 18, Number 2

Science Programs for an Antarctic Infrared Telescope

Having established that the Antarctic plateau provides the best conditions for infrared astronomy on the Earth, it is necessary to consider the demands of the site before determining what programs might best be pursued there. Clearly, constructing 8m size telescopes in Antarctica is a formidable challenge (though easier than facing the same challenge in space), and the development of Antarctic astronomy will be through smaller facilities first. These will establish the necessary infrastructure, as well as provide test-beds for developing appropriate engineering technologies and operating practice, needed before large-scale facilities can be built. Indeed, through the operation of the SPIREX 60cm telescope (Hereld, Rauscher & Pernic, 1990) with the Abu infrared camera (Fowler et al. 1998), a prototype system has already demonstrated that infrared astronomers can successfully conduct significant scientific program from the South Pole over the Antarctic winter (Burton et al. 2000, Brooks et al. 2000).

In view of this necessary staged development program, consideration of proposed intermediate scale infrared facilities must pay regard to providing complementary functionality to the large (8m) telescopes elsewhere, and to exploiting areas where the Antarctic facility clearly provides the most powerful tool for the job. Although Table 1 provides a direct comparison of sensitivity, it must be stressed that the costs of the facilities being compared differ enormously, with a 2m Antarctic telescope projected to be over an order of magnitude cheaper than existing 8m facilities. In addition, instrumentation for a small telescope is much less expensive than instrumentation for a large telescope, because the Area-Omega product, A $\Omega$ , of the telescope propagates through the whole system. Giving regard to such costs, for subsequent developments beyond a 2m facility it may prove practical to consider developing a suite of such telescopes, each designed for a specific project with a tailored instrument. These telescopes may also be linked together to form an interferometer. The particularly stable level of the sky fluxes in the mid-IR window may make this an attractive proposition.

There are three particular areas where an intermediate size (2m class) Antarctic telescope can be superior to any other telescope on Earth:

Wide-field thermal infrared imaging. The reduced sky background and improved transmission allow a 2m size Antarctic facility to be at least as sensitive as a temperate 8m class facility for imaging. Moreover, the Antarctic telescope can survey large areas of sky rapidly because of its wider field of view. The telescope, and the instrumentation, is vastly cheaper. Such a facility can complement the 8m telescopes by finding the sources for the larger facility to study in depth with high resolution spectrometers.
Continuous observation at 2.4 $\mu$ m, where the sky background is lowest, on sources which are always above the horizon. In the thermal infrared continuous observation is possible year-round, and is not just confined to dark periods.
Mid-IR interferometric imaging, exploiting both the greatly reduced background in the 8-14 $\mu$ m range, and the improved sky stability. Such an interferometer may also provide a test-bed for the ambitious space-based projects which aim to detect Earth-like planets around other stars.

We see at least five science programs where these advantages will enable significant advances to be made in our knowledge of the Universe:

Near-IR studies of the environment of embedded star forming complexes, imaging the molecular, neutral and ionized gas through their infrared spectral features.
Near- and mid-IR imaging of the embedded population of star forming regions, determining their complete population and in particular identifying the youngest members, and the incidence of disks around them.
Near-IR surveys for proto-galaxies and the early stages of star formation in galaxy evolution.
Micro-lensing studies of the stars towards the Galactic centre at 2.4 $\mu$ m, utilising the low sky background and high surface density of stars, in particular to identify the incidence of secondary lensing from planetary systems.
Mid-IR interferometric imaging of nearby star systems to search for proto-planetary disks, zodiacal dust clouds and Jovian-size planets around them.

We discuss these programs in more detail below. There are, of course, many other science programs which might be considered: for instance, continuous observations of variable sources, removing aliasing problems which can occur if sampling is interrupted every 24 hours.

The Environment of Star Forming Complexes

**Figure 3:** Infrared image of a 20' section of the molecular ridge of NGC6334, obtained with the 60cm SPIREX/Abu telescope at the South Pole (from Burton et al. 2000). It shows the extensive PAHs emission features at 3.3 $\,\mu$ marising from photodissociation regions surrounding five sites of massive star formation. The electronic version of this picture, obtainable from the journal website, is in 3 colours and shows the photodissociation region surrounding bubbles of ionized gas and embedded sources (blue: 3.3 $\,\mu$ m PAH, green: 3.5 $\,\mu$ m L-band, red: 4.05 $\,\mu$ m Br $\alpha$ ).
$\begin{figure} \vspace{-5cm} \begin{center} \psfig{file=6334_pahscolour.ps,height=25cm}\par\end{center}\end{figure}$

While massive star formation is one of the most spectacular events in the Galaxy, paradoxically it is poorly understood. This is because of both the short timescales for the various stages of the process, and because of the many interacting phenomena for which it is hard to disentangle cause and effect. The environment of such star forming complexes, which dominate the southern Galactic plane, can be studied in the thermal infrared through the spectral features from ionized, neutral and molecular species that are present. HII and ultra-compact HII regions can be traced in the Br $\alpha$ 4.05 $\mu$ m line, even when deeply embedded. Polycyclic Aromatic Hydrocarbons (PAHs), organic molecules that are fluoresced by far-UV radiation from the young stars and trace the edge of photodissociation regions, are visible through a spectral feature at 3.3 $\mu$ m. They can be imaged at high spatial resolution, unlike other prime tracers of these regions, such as the far-IR [CII] 158 $\mu$ m line. Excited H₂ emission, resulting from either shocks or UV-fluoresence, can be imaged in the v=1-0 Q-branch lines at 2.4 $\mu$ m, which are both stronger and suffer less extinction than the commonly used 1-0 S(1) line at 2.12 $\mu$ m. Several solid state absorption features are also present, for instance the ice band at 3.1 $\mu$ m.

As an example of the potential for this kind of study, Fig. 3 shows an

18' x 18' region of the star forming complex NGC6334, observed with the SPIREX/Abu camera from the South Pole (Burton et al. 2000), in the PAH and Br $\alpha$ features, as well as in the L-band continuum at 3.5 $\mu$ m. The pixel scale in this image is 0.5'', and combining the 1.5'' diffraction limit with 1 hour of unguided tracking, the typical resolution achieved was $\sim 3''$ . Shells of photodissociated gas surround bubbles of ionized gas in which embedded, massive protostars reside. Despite the modest size of the SPIREX telescope (just 60cm), these are the deepest images yet obtained at these wavelengths at this spatial resolution. The small aperture, however, also made possible the wide field of view with a similarly modest instrument.

Complete Population Census of Star Forming Regions

A key goal for studies of star formation is to undertake a complete population census of star forming clouds in order to determine the number and types of stars that form in them, and how this varies between different complexes. To do so requires observations in the thermal infrared (

$\lambda > 3\,\mu$ m). These wavelengths not only penetrate to the depths of cloud cores, but also allow us to distinguish between the embedded population and background stars. In simple terms, young stellar objects are surrounded by warm (few hundred K) disks which emit strongly at

$\lambda > 3\,\mu$ m, and thus are readily distinguished in infrared colour-colour diagrams (e.g. [

$\rm 1.65\mu m-2.2\mu m$ ]/[

$\rm 2.2\mu m-3.8\mu m$ ]) from reddened stars. Near-IR colour-colour diagrams (e.g. [

$\rm 1.25\mu m-1.65\mu m$ ]/[

$\rm 1.65\mu m-2.2\mu m$ ]), while relatively easy to construct because of the better sensitivities available, show only small IR excesses from the disks. These excesses are readily confused with reddening, and the surveys fail to identify the most deeply embedded sources.

The problem has been that at 3.8 $\mu$ m sensitivities are typically 4-5 magnitudes worse than at 2.2 $\mu$ m from most observing sites, thus limiting the work that has been done in this waveband. Needed are deep, wide-field surveys of comparable sensitivity to those conducted at 2.2 $\mu$ m in order to determine the complete stellar membership of a star formation region. Such an opportunity is afforded by an Antarctic telescope through the greatly reduced thermal background at these wavelengths over temperate sites.

Brown dwarfs--cool sub-stellar objects--may also be identified through the deep absorption band at 3.4 $\mu$ m, using narrow band filters on and off the band to determine ``colours". Even cooler protostellar objects would be detectable in the mid-IR, for instance embedded sources within `hot molecular cores' (e.g. Walsh et al. 2001), suspected of being the first stage in the process of massive star formation. Imaging through narrow band (1 $\mu$ m wide) filters at 8.5, 10.5, and 12.5 $\mu$ m, where the background is at a minimum in the mid-IR window, will allow determination of spectral colours of these cooler objects, and thus help to place their evolutionary state.

Protogalaxies and the First Star Formation

The star formation history of the Universe is being probed through deep pencil-beam surveys, of which the Hubble Deep Fields (HDF, Williams et al. 1996) are the most prominent examples. At the faint end of the samples the relative number of peculiar or disturbed galaxies rises dramatically, suggesting that processes to do with star formation (e.g. mergers, starbursts) are active in these sources. However, these galaxies also correspond to the most distant in the samples, with the highest redshift, and in the visible the rest frame being imaged is that of the far-UV. Here star formation is not at its most apparent, and dust absorption can be significant. An Antarctic telescope can search extraordinarily deeply in the 2.4 $\mu$ m ``cosmological window'' to where, for example, the H $\alpha$ line is red-shifted at z=3. It could undertake the first high spatial resolution, wide-field surveys at 3.8 $\mu$ m (L-band), where the visible light from z=5 galaxies would be observed. While the magnitude limit of the HDF (I $\sim 28$ mags.) will remain far deeper than that which an Antarctic 2m telescope will reach at 3.8 $\mu$ m (L $\sim 19$ mags. in 24 hours), the colours of high-z galaxies are particularly red. For instance, an E/S0 galaxy at z=1.4 has an unreddened colour of

$\rm V - L \sim 10$ . Thus a galaxy with V=28 and L=19, barely detectable in the HDF, would be detectable with an Antarctic 2m telescope in a day of integration. Moreoever, redder and presumably more interesting galaxies, not seen in the HDF, would also be detectable.

Micro-lensing towards the Galactic Centre

Gravitational micro-lensing occurs if the geodesic from a star to us passes sufficiently close to a massive, foreground object that its path is bent, or lensed, splitting the light into multiple images (Paczynski 1986). If there is a planet near one of the images an additional lensing effect can occur (Gould & Loeb 1992). The amplitude and light curve of such an event depends on the geometry of the orbit and mass of the planet, but typically will cause a perturbation on the microlensing light curve with a magnitude of a few percent for a few hours. If there is a planet present in the lensing system the probability of detecting a lensing signature from it is reasonably high if the sampling is frequent and the photometric accuracy high (Albrow et al. 1999). To maximize the possibility of finding such events a dedicated telescope should continuously image the same region of sky where the stellar density is high. Nowhere is this more so than towards the Galactic centre. Furthermore, the Galactic centre becomes readily detectable at 2.4 $\mu$ m (extinction precludes observation at much shorter wavelengths), the very waveband where the sky background is lowest in Antarctica. Moreover, the Galactic centre is always visible from the South Pole. For example, a 2m telescope equipped with only a single 1024² array with 0.6'' pixels, mosaicing on a 4 x 4 grid, could image a

40' x 40' region roughly every 20 minutes, achieving a sensitivity of $\sim 17.5$ mags at 2.4 $\mu$ m. Towards the Galactic centre every pixel would contain at least one star! As calculated by Gould (Gould 1995), the optical depth for lensing is then unity; i.e. we would always expect to find at least one lensing event underway. Such a facility would be a powerful tool for exploring the incidence of planetary systems through the secondary lensing signature imposed on the micro-lensing light curve.

Interferometry of Proto-Stellar Disks and Jovian Planets

One of the great challenges facing astronomy, and the focus of major national programs such as NASA's Origins program, is the search for Earth-like planets. Several grand design projects have been envisaged towards this goal, for instance NASA's Terrestrial Planet Finder (Beichmann, Woolf & Lindensmith 1999) and ESA's Darwin (Penny et al. 1998). These are space-based nulling interferometers, a suite of telescopes operating in mid-infrared where the unfavourable contrast between star and planet is least. Such facilities are not likely to be built before the middle of the 21^st century, and many major technological issues remain to be addressed first. Several ground-based interferometers are now under construction, such as the Very Large Telescope, the Large Binocular Telescope and the Keck Telescopes, with the intermediate goal of imaging circumstellar disks, zodiacal dust and Jovian planets in nearby stellar systems. An Antarctic infrared interferometer (AII) is an obvious next step after a 2m class telescope, exploiting the reduced background, the improved sky stability compared to temperate sites, and the constant airmass of sources. We envisage the AII as a suite of 2m size telescopes, initially with just two connected interferometrically, but readily expanded for relatively low cost by the addition of more telescopes, to explore the optimal configuration for imaging other solar systems. It would provide the most powerful ground-based instrument for this purpose.