SALT HRS Public Pages
As previously mentioned, the spectrograph is designed with a range of observational programs in mind. For this reason the science drivers are fairly varied and include:
- stellar radial velocity measurements (stellar kinematics, Local Group, extra-solar planets);
- stellar atmosphere analysis (chromospheric activity, seismology, Doppler tomography);
- chemical compositions of stars (abundance analysis, isotopes, nuclear chronometry, SMC/LMC);
- interstellar and intergalactic absorption (diffuse interstellar bands, Lyman-alpha forest, high redshift absorbers)
For these reasons, the HRS has been developed with four different modes of operation: low-, medium- and high resolution modes, and a further high-stability mode. Overall this gives the instrument the following capability:
- resolving powers R = 16000, 37000 and 67000
- wavelength coverage from λ = 370-890nm
- high mechanical and thermal stability
- some limited multi-object capability
HRS is of dual beam, white pupil design. This means that the instrument has two channels, a 'red' channel (555-890mm) and a 'blue' channel (370-555nm) - meaning the optics, coatings and CCD detectors can be optimized in size, shape, cost and efficiency for different wavelengths. In the case of the HRS, a dichroic is used to separate the red and blue arms. The white pupil design is also featured on the UVES system at the European Southern Observatories (ESO) Very Large Telescopes (VLT). In such a design, a parabolic mirror is used as a collimator for both the echelle grating and the cross-dispersers, minimizing the size (and hence complexity and cost) of camera optics. The entrance slit light is collimated onto the echelle, which disperses light into a spectrum (the orders of which are superimoposed on one another). The collimator second pass then recombines this spectrum into a single pupil image (the 'white pupil') on the cross-dispersers, which are used to to separate the orders. In the HRS, the cross-dispersers are volume phase holographic gratings (VPHGs). Finally, camera systems images these cross-dispersed orders onto the detectors.
The echelle grating is an R4 grating (having a blaze angle of 76°), with 41.6 grooves/mm and is illuminated with a 200mm diameter beam. The instrument uses a single monolithic grating, divided in two. The small gap between the two sections is used to relay stray light (via a fibre feed) to a photon counting device, to be used as an exposure meter. The medium- and high-resolution modes also employ a fibre image slicing system.
The only moving components in the instrument are the shutters, camera focusing mechanisms and the fibre interchange mechanism. Resilience to temperature and pressure changes is ensured by enclosing all optics in a vacuum vessel at 2hPa. This is then insulated with a thermal blanket and housed in a temperature stabilized enclosure in the basement below the telescope.
A schematic of the design can be seen here in Figure 1:
In order to feed the instrument in the spectrograph room with light collected at the telescope, a 35m long optical fibre feed is used. At the telescope end, these are mounted at the SALT Fibre Instrument Feed (FIF), which accomodates 12 fibres in two rows of 6. 5 fibre pairs are for use with the HRS. The instrument uses (at the minimum) a thorium argon (ThAr) lamp for wavelength calibration and a white lamp for flat-fielding and echelle order definition. The input face of each fibre at the FIF cemented to a fused silica window, which has been multi-layer over-coated to reduce reflection losses entering the fibres (and hence the instrument).
The fibres terminate inside the vacuum vessel, where a moveable mask allows selection of the appropriate fibres. Three pairs of fibres from the FIF are used to feed the 'fixed sky and object' channel of operation - one pair for each of the low-, medium- and high resolutions. In this mode, the science object and a neighbouring patch of 'empty' sky are simultaneously observed. This is useful for background subtraction. Fibre for the low- and medium-resolution modes have a core size of 500 micrometres, with the high-resolution fibres featuring a 350 micrometre core diameter.
An additional pair of fibres is allocated for the 'nod and shuffle mode'. This channel will only be available in low-resolution mode, since at higher resolutions, the inter-order spacing on the detectors is prohibitively small for the shuffling to take place. It is an alternative method for background subtraction of sky emission, and should be particularly useful on SALT - given that background scales proportional to telescope diameter squared, and SALT is an 11m telescope.
The final fibre pair is to be used for the dedicated 'high stability mode'. This channel functions similarly to the high-resolution mode, but has been optimized for Doppler precision measurements. The mode is arranged such that there are no moving components, and with the additional benefit of a fibre double-scrambler and an iodine cell. Whilst these additions improve precision, there is of course a trade-off, in that throughput is reduced. For this reason the mode is added as an aditional standalone mode. The double scrambler is used to reduce spectral modal noise, by increasing the radial scrambling, and in doing so, removing beam 'memory'. This may prove of particular significance in implementation with the SALT telescope, which has a variable illumination and obscurations (see Figure 2).
Figure 2: The variable illumination of SALT as the object migrates across the mirror. This is due to the unique tracking style of the telescope.
The iodine cell is used to imprint a forrest of absorption lines in the observed spectra, making precise wavelength calibration possible. This is vital for ensuring precise Doppler spectroscopy results, useful amongst others, for extra-solar planet detection (see Figure 3).