An information note concerning the expected performance of groundbased 8m telescope optical and NIR spectrographs.
As part of the ASWG deliberations on the choice of multiplexed (optical-) NIR spectrograph for NGST, it is necessary to have a clear idea of the region of spatial/spectral/FOV parameter space where 8-10m (and larger) groundbased telescopes may become competitive with NGST. In general, for reasons of sky and detector background, this is expected to be at wavelengths less than about 2.5 microns and for higher spectral resolutions.
In order to help define this boundary as closely as possible, we have examined the actual and projected performance of the existing and planned instrument suite of the ESO VLT. While the VLT consists of four of a number of similarly sized telescopes which will be operating before and during the NGST era, we expect it to be broadly representative of their capabilities. It is important to note, however that - particularly at the longer wavelengths where thermal backgrounds become important - different telescopes and instrument suites may be optimised in different ways.
Comprehensive, but not entirely homogeneous descriptions of the VLT instrument suite are available at the ESO VLT instrument website. Of these, the instruments which perform spectroscopy at wavelengths less than 5 microns are:
The instrument summaries referred to above give simulated and, in some cases, measured values of instrument performance. We have reviewed these values and investigated (some of) the underlying assumptions and found them to be sufficiently inhomogeneous that drawing quantitative conclusions could be misleading. Having said that, the figure shows a set of limiting fluxes for spectroscopy covering 450 < R < 105 culled from the performance descriptions.
The NGST NIR-SPEC at R=1,000 is taken from the Yardstick ETC for point sources which assumes a detector dark current of 0.02e- s-1 (ie. this is a hypothetical instrument: the VLT measurements and calculations include a good measure of current reality!). The VLT calculations generally assume the dark current to be 0.1 which has a significant effect on the high R CRIRES prediction. The VLT instruments with R ranging up to a few thousand make different assumptions about image quality but the ISAAC figures do not include AO and are measured under typical observing conditions with seeing ~ 0".6. For R greater than about two to three thousand, it is possible to see a substantial fraction of the shorter wavelength NIR spectrum between the OH lines where the sky is very dark. A groundbased instrument with a dark current of around 0.1e- s-1becomes detector noise limited at R~3,000 between the OH lines when looking more than about 30 degrees from the bright Moon. The figures for the SINFONI instrument appear better than ISAAC at R=500 because it observes at a higher R between the OH and does software binning of the data back to low resolution. At its higher resolutions, ISAAC is able to operate in a similar fashion. The differences between space and groundbased sensitivity below about 2.5 micron then depends primarily on the ability to exploit the diffraction limit of the telescope.
The bottom line is that for a diffraction limited telescope at wavelengths of less than about 2.5 microns at R ~ 3,000, a substantial fraction of the spectrum between the OH emission can be observed from the ground with a sensitivity similar to NGST.
The two major VLT MOS/IFS instruments (VIRMOS) operate in natural seeing and have a large accessible FOV (~ 200 arcmin2 for MOS and 40 x 40 elements for IFU). They can, to some extent, exploit the dark sky between the OH lines in the NIR and so will, like NGST, be dependent on the detector noise characteristics. They will not, however, be fed with diffraction limited images. An instrument like CRIRES, working at very high R, will be almost entirely dependent on detector noise (and the ability to exclude stray light from the instrument), only becoming sensitive to sky and telescope background at its longer wavelengths. The only currently planned instruments which will use AO to exploit diffraction limited, low/intermediate resolution spectroscopy are CONICA/NAOS, which has a maximum FOV of around 1 arcmin2 with 100mas sampling, and SINFONI with a FOV variable from 1.25 to 8.75 arcsec on a side.
The regions where groundbased spectrographs can most closely approach NGST are clearly at the shorter wavelengths and at higher spectral resolutions. Where, exactly, to draw the boundary depends on the details of the observation, the instrument and, most particularly on the detector noise performance. It is clear, however, that even at resolutions as low as a few thousand, groundbased instruments can exploit the dark sky between the OH emission with an efficiency with increases with increasing R (see the study by Matthew Colless and Kieth Taylor for AUSTRALIS. The relevant section is the scientific case). This may be most directly relevant to the DRM programmes which ask for R ~ 3,000 for kinematic studies of galaxies: this programme is already being attempted with ISAAC for objects with redshifts which put appropriate emission lines in OH gaps. 'Astrophysical' spectroscopy with R <~ 1,000 remains firmly in the NGST domain.
We thank Alan Moorwood and others at ESO for providing information about the VLT instrument suite and Alan, in particular, for calculating NGST/groundbased limiting efficiency ratios using a common set of assumptions and real measurements from Paranal of the sky brightness at J and H between the OH emission.
Email from Marcia Rieke on 29 Sep. 1999 with attached list of instrument questions from Simon Lilly.
See: Gemini Preprint #21, F. Gillett and M. Mountain,
On the Comparative Performance of an 8m NGST and a Ground Based 8m Optical/IR Telescope.
available at http://www.gemini.edu/documentation/preprint.html
| 7 October 1999 | Bob Fosbury/Stefano Cristiani | Send Comments | NGST home |
|---|
