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

Next Section: Science Programs for an
Title/Abstract Page: Science Goals for Antarctic
Previous Section: Introduction
Contents Page: Volume 18, Number 2

Subsections

Results from Site Testing

Infrared Sky Brightness

From the extensive site testing program undertaken at the South Pole the following characteristics of the infrared background have been determined:

Sky background in the K-dark window (2.27-2.45 $\mu$ m) as low as $\sim 100 \mu$ Jy arcsec^-2 (20-100 times less than at temperate sites) (Ashley et al. 1996, Nguyen et al. 1996).
Sky background in the L-band (3-3.8 $\mu$ m) $\sim 100$ mJy arcsec^-2 ( $\sim 20$ times less than at temperate sites) (Phillips et al. 1999).
Sky background in the N-band (8-14 $\mu$ m) as low as $\sim 20$ Jy arcsec^-2 ( $\sim 20$ times less than at good temperate sites)(Chamberlain et al. 2000).

For background limited imaging of extended regions a 2m telescope requires a 16-fold reduction in background if it is to achieve the same sensitivity as an 8m temperate-latitude telescope. It has improved sensitivity for larger background reductions, or if the lower background is accompanied by superior atmospheric transmission.

This comparison is quantified further in Table 1, which shows the relative signal to noise ratios for performance limited by the sky background, obtained for observations using four different telescopes: an 8m telescope on Mauna Kea (ie. Gemini); a 4m at Siding Spring Observatory in Australia (ie. the Anglo Australian Telescope), a 2m on the Antarctic plateau (ie. the proposed Douglas Mawson Telescope), and an 8m telescope in Antarctica. The comparison is for broad-band imaging at three representative wavelengths; K-band (2.2 $\,\mu$ m), L-band (3.65 $\,\mu$ m) and N-band (11.5 $\,\mu$ m) (for K-band the performance of an Antarctic telescope at 2.37 $\,\mu$ m is compared to that of a temperate-latitude telescope at 2.15 $\,\mu$ m, where airglow emission dominates the background). Two cases are shown: (i) wide-field imaging, used for extended objects, where the pixel size (in arcsec) is taken to be the same in each case, and (ii) point-source imaging, where diffraction limited performance is assumed to be achieved by each telescope. The S/N ratio is proportional to $(\rm D \eta / \theta) S^{-0.5}$ , where D is the diameter of the primary, $\theta$ is the spatial resolution, $\eta$ is the atmospheric transmission and S is the sky background at that waveband. Performance comparisons have been normalised to the Mauna Kea telescope, all other factors being taken as equal.

Also shown in Table 1 are achievable sensitivities¹ in magnitudes, for a $5 \sigma$ detection in 1 hour, taking into account telescope emission and system performance of the telescope + instrument + detector. By taking the same instrumental parameters at each site, this allows a direct comparison of the performance achievable as a result of the site conditions and telescope aperture. We note that for Siding Spring, in thermal wavebands, and in N-band at all sites, telescope emission is comparable to, or slightly greater than, the sky emission. The S/N relation above does not strictly hold in these cases and thus the sensitivities listed above (which have included telescope emission) are correspondingly slightly worse than the S/N numbers above would indicate. The S/N relation also does not hold for the comparison in the K-band, where the optimal observing wavelengths and bandpasses would be slightly different between Antarctic and temperate sites. Also listed in the table is the sky background in each waveband, for each site, in Jy/sq. arcsecond.

**Table:** Relative S/N Ratios and Sensitivities for Different Telescopes
Telescope	Mauna Kea 8m		SSO 3.9m		Antarctic 2m		Antarctic 8m
	Gemini		AAT		DMT
	Wide	Point	Wide	Point	Wide	Point	Wide	Point
Waveband	Field	Source	Field	Source	Field	Source	Field	Source
K	3 x 10^-3		3 x 10^-3		1.5 x 10^-4		1.5 x 10^-4		Background
(2.15 $\,\mu$ m vs	1.0	1.0	0.5	0.2	1.1	0.3	4.4	4.0	Relative S/N
2.37 $\,\mu$ m)	21.5	23.1	20.5	21.3	21.2	21.0	22.8	24.1	Sensitivity
L	2		3		0.1		0.1		Background
(3.65 $\,\mu$ m)	1.0	1.0	0.4	0.2	1.1	0.3	4.5	4.5	Relative S/N
	16.7	17.8	15.3	15.5	16.9	16.4	18.4	19.4	Sensitivity
N	200		1000		20		20		Background
(11.5 $\,\mu$ m)	1.0	1.0	0.2	0.1	0.8	0.2	3.2	3.2	Relative S/N
	11.8	11.5	9.9	8.9	11.2	9.5	12.7	12.5	Sensitivity

Sky backgrounds (in Jy/sq. arcsecond), relative signal-to-noise ratios, and sensitivities in magnitudes ( $5 \sigma$ , 1 hour) comparing 4 telescopes in K, L and N bands for both wide-field (per square arcsecond) and point-source (ie diffraction-limited) imaging. For full details see the text.

For wide-field imaging an Antarctic 2m has similar sensitivity in the thermal infrared to that of an 8m telescope on a good infrared site, such as Mauna Kea, but has potentially a much wider field-of-view (as well as the opportunity to devote substantial time allocations to specific projects). Both these telescopes have gains of 2-5 times over current 4m class telescopes. If diffraction-limited imaging with an 8m is achieved then these temperate-latitude telescopes are superior for point-source imaging. However, if an Antarctic 8m were to be built it would be 3-5 times more sensitive than a temperate-latitude 8m, for all types of observation.

Seeing

The ice-level seeing at the South Pole is relatively poor ( $\sim 1.5''$ ). However the degradation occurs almost entirely in the lowest $\sim 200$ m of the atmosphere, with the free seeing above the boundary layer believed to be $\sim 0.3''$ (Marks et al. 1996, 1999). For a telescope at ice-level, adaptive optics will be able to remove much of the seeing. The correction of the boundary layer contribution will result in an isoplanatic angle of $\sim 60''$ (Marks 2001) at 5000Å, $\sim 30$ times greater than achievable on sites like Mauna Kea. This is possible because the seeing is caused by a narrow layer close to the ground, rather than having some components arising from a high-altitude jet stream. For infrared observations virtually all the sky will have a star brighter than $m_{\rm K} =10$ within the isoplanatic patch, and therefore be suitable for adaptive optic correction. Because of the uniquely favourable conditions on the Antarctica plateau, multi-conjugate adaptive optics (MCAO, e.g. Rigaut, Ellerbroek & Flicher, 2000), with its immense complexity, should be completely unnecessary.

On the summit of the high plateau, there is reason to believe that, based on measurements of the temperature inversion close to ice-level at several locations on the plateau (e.g. see Scwerdtfeger 1984), the boundary layer will be confined to an even lower altitude than at the Pole. The seeing within the boundary layer depends on the wind shear within the boundary layer and on fluctuations in the vertical temperature gradient (Marks 2001). Wind shear is minimised on the summit of the plateau where the slope of the ground is zero. A tower might be built to raise a telescope above the boundary layer on a high plateau site. However, even without a tower, the decreased height of the inversion layer improves the prospects for adaptive optics corrections still further, as it increases the isoplanatic angle. The AASTO program, which is currently site-testing high plateau locations (see, for example, Storey, Ashley & Burton, 2000), is aimed at quantifying such issues.

Model Calculations for Atmospheric Emission and Transmission

**Figure 1:** The measured near-IR sky spectrum (Phillips et al. 1999) (from 1.5-2.5 $\mu$ m and 2.9-4.1 $\mu$ m) and mid-IR sky spectrum (Chamberlain et al. 2000) (5-14 $\mu$ m) at the South Pole. Over-plotted, with a dashed line, is a model spectrum, corresponding to 164 $\mu$ m of precipitable H₂O, plus an aerosol visibility of 100km. The model fails below 2.3 $\mu$ m because of the neglect of airglow emission.
$\begin{figure} \begin{center} \psfig{file=spie_00_fig1.ps,height=9cm}\end{center}\end{figure}$

**Figure 2:** Model calculation of the atmospheric transmission at the South Pole across the thermal infrared, from 2-500 $\,\mu$ m, corresponding to 164 $\mu$ m of precipitable H₂O, plus an aerosol visibility of 100km (see Hidas et al. 2000).
$\begin{figure} \begin{center} \psfig{file=trans_164um_100km.ps,height=9cm}\end{center}\end{figure}$

In Fig. 1 are shown measured sky spectra from the South Pole, from 2 to 14 $\mu$ m, overlaid with a model fit using the atmospheric modelling program MODTRAN (see AFRL/VSBM). From the parameters to the fit a typical value of 164 $\mu$ m for the precipitable water vapour, with an aerosol visibility of 100km, was obtained (Hidas et al. 2000). This has permitted estimates to be made of the sky background and transparency for wavelengths not covered in the site testing programs. Fig. 2 shows the model transmission calculated for these parameters. New windows for ground based astronomy are opened in the mid-IR between 20 and 50 $\mu$ m, and even windows at 200 and 220 $\mu$ m may be accessible.

Next Section: Science Programs for an
Title/Abstract Page: Science Goals for Antarctic
Previous Section: Introduction
Contents Page: Volume 18, Number 2