To answer the increasingly challenging questions of modern astronomy, larger and larger optical telescopes have been built in the last few decades, equipped with a sizable suite of focal-plane instruments devoted to observations at all wavelengths, and spatial and spectral resolutions. As a general rule, for a given spectral resolution and under seeing-limited conditions, larger telescopes require larger instruments because spectrograph dimensions scale with the slit width, which scales with telescope diameter.
For example, the High Resolution Echelle Spectrometer at the Keck telescope1 has a 1.2 × 0.3m2 mosaic grating consisting of three segments, and collimator and camera mirrors of the same order of magnitude. Some attempts to reduce the dimensions of the optics have been made, for instance for the Ultraviolet and Visual Echelle Spectrograph2 at the Very Large Telescope (VLT), where a smaller 0.8 × 0.2m2grating has been used.
If we were to linearly scale such instruments to 30–40m-class telescopes, such as the 42m European Extremely Large Telescope3 (E-ELT, illustrated in Figure 1), this would result in three to five times larger components and thus very large, heavy, and expensive instruments. Moreover, manufacturing constraints have been reached in many areas. For example, the ruling engines that are used to make diffraction gratings—the main dispersing component of a spectrograph—are limited to ruled areas of around 40 × 30cm2.
Figure 1. Model of the E-ELT within its enclosure. The main mirror has a diameter of 42m. The COsmic Dynamics EXperiment (CODEX) instrument will be situated at a stationary location below the telescope and receive its light via an optical train of mirrors and fibers. (Courtesy: European Southern Observatory.)
We have investigated a new design approach4 that uses slicing techniques (at image or pupil planes) and ‘anamorphic’ magnification at different levels to keep instrument size and cost affordable. We can use these methods because the resolving power of a spectrograph is given by the length of the grating only. Because its width does not play any role, if one slices the pupil into two (or more) subpupils superimposed on the same grating, the grating length can be reduced by the same factor. The price one pays is an increase in detector area, because the images from the two pupil slices cannot be superimposed (preservation of the etendue).
A second technique we explored is beam compression along the grooves before it reaches the grating, to reduce the grating width. For example, a 1:4 anamorphic compression will reduce the grating width by a factor of four, and significantly decreases the number of segments to be used in a grating mosaic. Such beam compression will require an even larger detector area, unless the beam is expanded again before it reaches the detector. Keeping in mind that spectral resolution does not depend on grating width, we can use anamorphism as a free parameter in spectrograph optimization, without affecting spectral resolution.
These two approaches have been used in the conceptual design of two instruments, one5 for the combined focus of the VLT, which is equivalent to a 16m telescope, and another—the COsmic Dynamics EXperiment (CODEX)6—for the 42m E-ELT. Both spectrographs share many design solutions, although different choices have been made to meet scientific requirements. Both are seeing-limited cross-dispersed echelle spectrographs, with a spectral coverage of 400–800nm and a resolution greater than 100,000.
Our designs have non-traditional optics7 before the entrance slit that introduce both pupil slicing and anamorphic beam compression. As usual in such instruments, the cross-dispersed format is generated by two dispersers operating in perpendicular directions. The main disperser is an echelle grating—working at high diffraction orders—while the cross-disperser can be a prism or a low-order grating.
The new designs use the cross-disperser just before the camera as a beam expander to create a more symmetric pupil onto the camera optics and reduce the detector area. A new component is under investigation that efficiently combines the functions of providing cross-dispersion and beam expansion: a ‘slanted-fringe’ volume-phase holographic grating.
These design techniques allow us to build high-resolution spectrographs for the next generation of extremely large telescopes with component sizes that are within present manufacturing capabilities and realistic budget envelopes. In the next year, the proposed instrument concepts will be further investigated, and we expect that some key components (such as the anamorphic pupil slicer) will be prototyped quite soon.
National Institute for Astrophysics
Brera Astronomical Observatory
Paolo Spanò is an optical engineer who works mainly in the field of high-resolution spectrometers for ground-based telescopes. He is also involved in space projects for European Space Agency missions.
Hans Dekker, Bernard Delabre, Gerardo Avila
European Southern Observatory
Garching bei München, Germany
Hans Dekker is a systems engineer and project manager for astronomical instruments.
Bernard Delabre is an optical designer of astronomical instruments and telescopes.
Gerardo Avila is an optical engineer and specialist in the field of astronomical applications of optical fibers.