Spectroscopy is a key technique in astronomy revealing such things as the cosmological red-shift (inferred distance), hidden matter in galaxies, how matter is devoured and blown out by suppermassive blackholes, and tells us the orbits of stars and their chemical abundances leading to the untangling of the history of the construction of galaxies via "galactic archaelogy". In its highly-multiplexed form (our speciality), it will (must!) tell us the Equation of State of the universe so that we can work out what most of the universe is actually made of! And, of course it allows us to detect exo-planets by seeing how the parent stars wobble.
The heart of a spectrograph (a.k.a. spectrometer) is a disperser that splits the light according to its wavelength so that a spectrum of each object (or part thereof) can be recorded by a suitable camera. There are three main technologies of interest to us:
(a) Surface relief gratings (SRG) replicated or carved out of glass which can be used in reflection or transmission. These "diffraction gratings" contain thousands of parallel grooves perpendicular to the direction of dispersion. The grooves are faceted with angles to maximise their efficiency at certain wavelengths ("blazing") and they can be combined with prisms to optimise their use in transmission.
(b) Volume phase holographic gratings (VPHG) imprinted in a photosensitive medium by holography. They work in the same way as SRGs but replace the grooves with harmonic variations in refractive index. They too can be blazed and sometimes perform better than the SRGs.
(c) everything else including miniaturised integrated photonic spectrographs which use arrays of waveguides to do the job (see the Astrophotonics pages) and energy sensitive (non-dispersive) detectors (e.g. Superconducting Tunnel Junction devices) which are far too expensive for us to play around with. Don't forget the humble prism which is good for low resolution and doesn't waste light into different orders like every other dispersive device
The big problem to solve: the bigger the telescope the bigger the disperser (if everything else like projected slitwidth is the same) so Extremely Large Telescopes need metre-sized gratings (unless you're smart) - but no one can make them - yet!
Read the link below for more information. If you must rip it off for your lecture course, please let me know and reference its author.
- Spectroscopy theory (last modified: 16 March 2010)