The combination of large telescopes and IR arrays with small pixels requires that IR spectrometers with a reasonably high throughput (i.e. with few pixels sampling a slit) must have very fast cameras. For example, an instrument equipped with the Rockwell 10242 array (18.5 m pixel size) and intended to work on a 8m class telescope with a slit sampled by 4 pixels must have a F/2 camera, while a much more extreme F/1.4 system is necessary to obtain 3 pixels/arcsec. The results presented here indicate that these types of cameras are relatively easy to design when BaF2-SF pairs are adopted.
Figure 3: Layout and ray tracing of F/2 and F/1.4 near-IR cameras. Spot diagrams and encircled energies are displayed in Figs. 4, 5 while the details of the design are listed in Table 2
Figure 4: Monochromatic spot diagrams (neglecting lateral chromatism) at several wavelengths and distances from the array center for the BaF2-SF6 cameras (Fig. 3 (click here)), the array is a 10242 with pixels of 18.5 m. The size of one pixel is indicated by the thick squares, while the larger squares are 2 pixels
Figure 5: Diameters of 80% and 90% encircled energies (in m) for the F/2 and F/1.4 cameras. The value is the maximum image size at the wavelengths of Fig. 4, for SF56A the curve at 2.5 m is shown separately. Note that BaF2-SF6 is excellent over the full range (and is in practice equivalent to BaF2-IRG2). The other pair is slightly better at the shortest 's and is achromatic up to about 2.2 m, but rapidly deteriorates beyond
Representative solutions for F/2 and F/1.4 systems are shown in Fig. 3 (click here), the spot diagrams and encircled energies are displayed in Figs. 4 (click here), 5 (click here) while the details of the lenses are listed in Table 2. In these designs we assumed a "typical'' 100 mm collimated (parallel) beam allowing for about 200 mm free space between the pupil image and the first optical surface. Note that the cameras were not optimized for lateral chromatism because this aberration does not influence the spectrograph performances.
The performances of the BaF2-SF6 cameras are remarkably good and virtually identical to those obtained using BaF2-IRG2. The BaF2-SF56A combination is also very good for m but not at longer wavelengths where their chromatisms decouple, but this limitation is not important for applications to instruments not extending into the K atmospheric window. As already mentioned in Sect. 2, the image quality is not influenced by the exact behaviour of with wavelength, and any reasonable uncertainty on this parameter can be recovered by simply refocusing the array.
The design of the cameras was somewhat simplified by the use of one aspheric surface on BaF2, but this should not be a problem because this material can be quite easily diamond turned. Another potential limitation is the size of the BaF2 elements, especially in the F/1.4 camera which requires lenses with mm useful diameters. However, these are not much larger than the 160 mm BaF2 aspheric collimator mounted in ISAAC, the ESO infrared instrument for the ESO-VLT 8m telescope (Delabre 1993). We believe therefore that the designs presented here are within the present possibilities of lens manufacturers.
|F/2 camera||F/1.4 camera|
|Surface||Radius (mm)||Thickness (mm)||material||Surface||Radius (mm)||Thickness (mm)||material|
|Lens 1||91.6531||40.00||BaF2||Lens 1||-1843.86b||30.45||BaF2|
|Lens 1||1944.23||4.761||Lens 1||-248.078||26.28|
|Lens 2||605.277||7.823||SF6||Lens 2||-489.278||10.00||SF6|
|Lens 2||167.455||3.000||Lens 2||345.473||29.56|
|Lens 3||116.127||26.62||BaF2||Lens 3||-4226.46||49.88||BaF2|
|Lens 3||152.138a||145.8||Lens 3||-127.895||85.98|
|Lens 4||57.9477||40.00||SF6||Lens 4||178.691||50.00||BaF2|
|Lens 4||48.2670||15.00||Lens 4||-348.386||71.84|