GHRS Instrument Handbook
Instrument Summary -- Why Use the GHRS?
Here we provide a brief overview of the basic properties of the GHRS. Each of these aspects is described in more detail in the next two chapters. Chapters 6, 7, and 8 provide illustrations of the GHRS and tables of instrument parameters.
The GHRS can obtain spectra from about 1150 to 3200 Å. These limits are set by the magnesium fluoride coatings on HST's optics and by the nature of the detectors. The additional two reflections introduced by COSTAR's mirrors significantly reduce throughput at the very shortest wavelengths (i.e., below Lyman-a) so that even very bright stars (e.g., m Col) have failed to produce detectable flux below 1150 Å. It is possible to observe bright stars out to 3400 Å.
With Side 1, observations may be made from 1150 to 1800 Å at R xbb 2,000, 25,000, and 80,000 (gratings G140L, G140M, and Ech-A, respectively). With Side 2, the options are R xbb 25,000 from 1150 to 3200 Å (G160M, G200M, and G270M) and R xbb 80,000 from 1680 to 3400 Å (Ech-B). For certain applications it can be advantageous to use grating G270M to wavelengths as low as 2100 Å because of its higher efficiency.
Routine calibrations on standard stars provide flux-calibrated spectra that are accurate to 10%. Relative fluxes obtained at different wavelengths should be good to better than 5%. The repeatability of fluxes is even better, being better than 1%; i.e., it is possible to compare measures of the same wavelength in the same star at different times to within 1% for observations with the LSA.*1
Within a single bandpass (i.e., one grating setting), relative photometric precision is limited by photon statistics for S/N < 30 and by detector non-uniformities above that, provided that the detectors are being used within the linear portion of their response. With suitable observing strategies, it is possible to achieve relative S/N as high as 900 (Lambert et al. 1994)*2.
We have found that the photometric sensitivity of the GHRS has not changed with time to within 1% or less. Some preliminary evidence suggests that the sensitivity may be dropping slightly below 1200 Å, but the effect there is no more than about 10%.
See Chapter 5 for a more detailed discussion of GHRS calibrations.
The source to be observed may be placed in a Large Science Aperture (LSA) or a Small Science Aperture (SSA). Because of the installation of COSTAR, the LSA is 1.74 arcsec square and the SSA is 0.22 arcsec square, although they retain their pre-COSTAR names (2.0 and 0.25, respectively). The LSA allows virtually all the light of a point source to pass (about 95%, depending on wavelength), and that fraction is insensitive to the precise centering of a star. As a result the LSA is the best choice for obtaining reliable fluxes. However, a star can drift some in the LSA, causing some degradation of spectroscopic resolution. This means that the SSA is better for obtaining spectra with the best resolution. The SSA has about 50 to 70% of the throughput of the LSA; using the LSA will degrade resolution by 10 to 20% compared to the SSA because of the wings to the instrumental profile, plus any effects due to smearing.
The LSA has a shutter which automatically closes when an observation with the SSA is being performed, in order to reduce stray light. The LSA is preferred when reliable fluxes must be measured and the SSA is better when a narrow instrumental profile is needed (for line profile work, radial velocities, etc.).
Most observers use the GHRS to accumulate photons for the time needed to reach the signal-to-noise they desire. In ACCUM mode the exposures may be as short as 0.2 seconds, although use of standard procedures for improving S/N usually limits exposures to no shorter than about 27.2 seconds (see ``Standard Patterns for Substepping and Background Measurement'' on page 112). The GHRS has a rapid readout mode (RAPID) that can obtain spectra as often as every 50 milliseconds, but that can only be done by sacrificing many features that are important for producing high-quality spectra. Some details on these observing mode are provided in Chapter 4. Note that relative timing of spectra, especially in RAPID mode, can be established quite accurately, but that absolute timing is much more difficult to determine because the spacecraft's clock records in units of 1/8 second. If you need very precise times for your observations you should consult STScI.
The limited availability of memory on the HST spacecraft means that there exists a maximum number of operating commands that can be in place for a single set of observations. That can be a limit for use of the GHRS in certain cases, described later (``Memory Usage'' on page 52).