The design and performance of high resolution échelle spectrographs in astronomy.

The design and performance of several high resolution spectrographs for use in astronomy will be described. After a basic outline of the required theory, the design and performance of HERCULES will be presented. HERCULES is an R2 spectrograph fibre-fed from the MJUO 1-m telescope. The échelle grating has 31.6 grooves/mm and it uses a BK7 prism with a 50° apex angle in double-pass for cross-dispersion. A folded Schmidt camera is used for imaging. With a detector having an area 50 x 50 mm, and pixels less than 25 µm, HERCULES is capable of resolving powers of 40,000 to 80,000 and wavelength coverage from 380 to 880 nm. The total throughput (from the fibre entrance to the CCD) is expected to be nearly 20% (in 1" seeing). Measured efficiencies are only slightly less than this. HERCULES is also shown to be capable of excellent radial velocity precision with no apparent difference between long-term and short-term stability. Several significant upgrade options are also described. As part of the evolution of the design of a high resolution spectrograph for SALT, several instruments were developed for 10-metre class telescopes. Early designs, based in part on the successful HERCULES design, did not meet the requirements of a number of potential users, due in particular to the limited ability to inter-leave object and sky orders. This resulted in the design of SALT HRS R2 which uses a mosaic of two 308 x 413 mm R2 échelle gratings with 87 grooves/mm. Cross-dispersion is achieved with a pair of large 40° apex angle BK7 prisms used in double-pass. The échelle grating accepts a 365-mm collimated beam. The camera is a catadioptric system having a 1.2-m primary mirror and three lenses made of BK7 each around 850 mm in diameter. Complete unvignetted (except by the CCD obstruction) wavelength coverage from 370nm to 890nm is possible on a mosaic of three 2k by 4k CCDS with 15 µm pixels. A maximum resolving power of R ≈ 80,000 is possible. For immunity to atmospheric pressure and temperature changes the entire spectrograph is designed to be housed inside either a helium atmosphere or a light vacuum. The spectrograph chamber is nearly seven metres long. An alternative to the R2 SALT HRS is also described. This instrument is an R4 dual beam spectrograph based on a white pupil layout. The design is based on suggestions by B. Delabre and follows closely this authors SOAR HRS instrument. SALT HRS R4 uses volume-phased holographic gratings for cross-dispersion and a 836 x 204 mm échelle grating with 41.6 grooves/mm. The grating will be replicated from two smaller gratings onto a single Zerodur blank. The spectrograph is split into blue and red arms by a dichroic located near the white pupil relay intermediate focus. Wavelengths from 370 nm to 890 nm are covered by two fixed format blue and red dedicated dioptric cameras. The detectors will be a single 2k by 4k CCD with 15 µm pixels for the blue camera and a 4k by 4k CCD with 15 µm pixels for the red. The size of the cameras is reduced significantly by white pupil demagnification from an initial 200-mm diameter collimated beam incident on the échelle grating to around 100 mm (in undispersed light) on the VPH gratings. The final SALT HRS R4 instrument is also designed to be immersed in a vacuum vessel which is considerably smaller than that proposed for the R2 spectrograph. SALT HRS R4 is currently being developed in detail and will be presented for a critical design review in 2005 April