Return to CRSP User Manual
This section presents spectra of the sky background obtained with CRSP on the KPNO 1.3-m (f/15), using slit 4 and all of the filters likely to be used with a particular grating. The intent is to provide a rough guide for estimating the maximum integration time possible for a given configuration and verifying that the wavelength setting is correct with reference to sky background lines. The plots are given in ADUs, and are all scaled to an integration time of 1 second. No correction has been made for filter efficiency. Because the 2.1-m and 4-m are also f/15, the background levels through the same slit should be unchanged, although the higher thermal emissivity of the 4-m will result in larger background than in the figures for wavelengths > 2.3 microns.
With grating 1, the sky background short of 2.3 microns consists almost entirely of emission lines of OH and O2. Where these lines are resolved, use of a wider slit will widen the lines, but have relatively little effect on the peak signal. Longward of 2.3 microns, or with the low resolution gratings, the background is essentially continuous, and the peak signal will scale with the slit width. The thermal background is subject to seasonal variations of a factor of two, and the OH lines may also vary by that amount, possibly during the course of a night. Also, the O2 0-0 line at 1.27 microns will decay over a few hours after the end of evening twilight.
Spectra with grating 1, and the L-band spectrum with grating 3, were assembled as a mosaic, so the wavelength vs. pixel relation is not linear over the spectrum; the plotted fit was obtained using the IRAF 'identify' routine. For an accurate listing of the OH wavelengths, see the wavelength calibration tables.
The final plot shows spectra of the zenith sky for solar elevation
angles of 1.6
The standard spectra are a composite of I,J,H,K, and L band spectra obtained with a 4 arcsec slit (3) on the 1.3-m telescope. They have been rescaled to units of ADU for 0.0 mag in 1s integration. The areal ratios of the 2.1-m and 4-m telescopes to the 1.3-m are 2.8 and 8.5, respectively. To first order, the standard spectra may be scaled by these factors to provide an estimate of the signal on the larger telescopes, keeping in mind that slit losses will increase on larger telescopes at constant resolution and that the spectrum will be spread over more rows.
Depending on the blaze and order in which a grating is used, the actual efficiency can vary significantly, even across a single band. This is illustrated by the following spectra, which were obtained from the standard spectra above through division by the dispersion. The results are plotted for the I, J, H, and K bands for gratings 2, 3, and 4. This format is useful in comparing expected performance for gratings of similar resolution (e.g., gratings 3 and 4 in the J band; 2 and 4 in the K band) for observations in which a particular region of the wavelength band is of primary importance.
The followiing are sky spectra in which the lines have been identified by wavelength (in vacuum Angstroms). The lines short of 2.3 microns are OH lines for which the wavelengths are well known (the wavelengths listed are the average of the two unresolved hyperfine components). Longer wavelength "lines" are generally bands of many lines (mostly H2O and CH4) which are blended at the CRSP resolution. Their wavelengths were determined by convolving solar FTS spectra to the CRSP resolution and measuring the effective wavelengths of well-determined features. The strength, shape, and effective wavelengths of these features may depend on temperature and atmospheric humidity, so they are recommended only for approximate dispersion solutions.