Gillett & Mountain (1998) made initial performance comparisons between
a cold space-based 8-m telescope and a ground-based 8-m facility with adaptive
optics. They emphasised the substantial advantages, because of lower backgrounds,
of the space-based facility. However they pointed out that, especially
for moderate to high-resolution spectroscopy, the advantage of the telescope
in space was very subject to assumptions about detector performance, and
at some wavelengths could vanish with a relatively small reduction in aperture.
Oschmann et al. (1999) developed this comparison to cover 20, 30,
50 and 100-m ELTs. They make conservative assumptions about imaging performance,
based on early experience with Adaptive Optics on Gemini: a falloff of
AO performance at short wavelengths is assumed, such that image size (50%
encircled energy diameter) is effectively 2.1 x (diffraction limit) at
1.0 um wavelength, scaling to 1.1 x (diffraction limit) at 2 um and longer
wavelengths. They adopted the detector performance characteristics, extrapolated
from present-day achievements, used by Gillett & Mountain (1998), and
assumed ``typical" Mauna Kea atmospheric properties. The performance of
the ground-based ELTs was compared with that of a (cold) 8-m telescope
in space, at 1AU from the sun, for imaging and for spectroscopy, assuming
10,000s total exposures assembled from realistic on-chip exposure times
in the face of cosmic ray hits and other limiting factors. The results
(from Mountain, 2000, personal communication) are summarised in Table 1.
|w'length (um)||Imaging (R=5)||Spectroscopy (R=104)|
These comparisons may be summarised as follows:
Table 1 shows that in the 1.25, 1.6 and 2.2 um windows the ELTs are at least competitive, and for the most part well ahead, in imaging performance. The huge gains with aperture for operations in the diffraction-limited regime are particularly apparent from the performance of the 100-m relative to the 20-m facility. It must be remembered too, that this performance goes hand in hand with a major advantage in spatial resolution.
In the near-IR (to 2.5 um wavelength) the ELTs are competitive or dominant in imaging sensitivity.
In the Thermal-IR (beyond 3 um wavelength)
ELTs are completely outperformed in sensitivity and (especially) speed
by the cold 8-m space telescope.
(Note that speed varies as
the square of the S/N).
It is immediately evident from Table 1 that all the ELTs dramatically outperform the space telescope in the 1.25, 1.6 and 2.2 um windows for spectroscopy: even the 20-m is an order of magnitude faster than the 8-m in space for spectroscopy at 2.2 um. The 50- and 100-m facilities outperform the space telescope at 3.5 um as well. Indeed the 100-m is competitive (i.e. not more than 10 x slower) for spectroscopy at all the wavelengths considered, while the 50-m delivers a useful fraction of the performance of the cold space 8-m even at wavelengths between 5 and 25 um. It must be noted, too ( c.f. Gillett & Mountain, 1998) that a ground-based facility has a natural advantage in keeping up to date in technological development. This means that the numbers in the Table will grow with time after launch of the space telescope; or, in fact, with lapse of time after its instrument design is frozen (unless the space facility occupies an orbit accessible to servicing missions and is designed to be upgraded).
In the near-IR (to 2.5 um wavelength) the ELTs dramatically outperform the 8-m space telescope for spectroscopy. (The larger apertures outperform the cold telescope out to 3.5 um)
In the Thermal IR (beyond 3 um wavelength) the
larger ELTs may be - or become - competitive with the 8-m
space telescope for spectroscopy in all the atmopheric windows.