Spiral B

Light from the AAT's f/8 Cassegrain focus enters a reimaging system which produces a magnified image of the telescope focal plane. A two-dimensional array of small lenses then samples the enlarged image (Fig.1), with each lens feeding the light to an optical fibre. Optical fibres allow the two-dimensional input array to be reformatted into a one-dimensional slit which feeds a purpose-built Littrow spectrograph. A spectrum is therefore produced for each lens at the input. An 18 m length of fibre is used to transfer the light from the Cassegrain to the spectrograph situated on the dome floor. Placing the spectrograph on the dome floor provides much better stability.

The phase A prototype version of SPIRAL was built jointly by the Institute of Astronomy in Cambridge, and the AAO, and was successfully commissioned in 1997 (Kenworthy 1997). SPIRAL A provided a field of view of approximately 3.5 arc seconds by 3 arc seconds by using an array of 37 small hexagonal lenslets (Parry 1997). Each lenslet is 4 mm across and corresponds to a sampling element of 0.5 arc seconds. The main difference between SPIRAL A and SPIRAL B is the field of view. SPIRAL B will use a microlens array with 32x16 1 mm square lenses. The sampling interval has also been increased to 0.7 arc seconds per lens so that the total field of view of SPIRAL B is 22 x 11 arc seconds squared.
 

Figure 1: Schematic of the operation of the SPIRAL integral field spectrograph.

SPIRAL B will use a crossed cylindrical microlens array. It consists of two arrays of cylindrical lenses which when placed together, with one array perpendicular to the other, form an array of square lenses. The array was custom made by Limo Gmbh in Germany. The microlens array is shown in Fig. 2. We must also position 512 optical fibres each at the focus of a microlens to an accuracy of about ±10 microns. This will be done with a specially made brass plate which contains an array of accurately drilled holes. The holes are difficult to produce as the drill must be only 200 microns in diameter! Fibres can then be inserted into the holes, the array polished, and the fibre array glued to the lens array. In total 10 km of optical fibre is needed for the entire fibre bundle. The total system throughput of SPIRAL B is expected to be around 12%. Predicted values of resolution (/), dispersion and wavelength coverage for SPIRAL B at a wavelength centre of 500nm:
 
 
 

Note that the dispersion assumes 30 micron pixels.

SPIRAL B is expected to be commissioned in late March 2000. More information about SPIRAL can be found at: http://www.ast.cam.ac.uk/~optics/spiral/spiral.htm or from the author, email: dl@aaoepp.aao.gov.au.

Fig. 2: A photograph of the SPIRAL B crossed cylindrical microlens array containing 512 1mm square lenses.

References