To obtain cameras with much larger numbers of pixels the general approach has been to construct "mosaics" of CCDs. As shown in Figure 7.13, CCDs that have been specially manufactured to ensure that three sides have no connections, and can therefore be butted very closely to one another, enable one to construct a larger mosaic. In this case, eight 2 K x 4 K chips were used to create a mosaic of 8 K x 8K. Even on the fourth side, connection space is kept as small as possible. The first operational wide-field CCD camera was the MegaCam instrument on the Canada-France-Hawaii Telescope (CFHT) which started science operations in 2002. Using 40 2 K x 4 K CCDs from e2v, this camera can image a field of view of 0.9 square degrees on the sky (Figure 7.14). Many other wide-field cameras are now in operation and more are planned. A useful metric to gauge the efficiency of a large-area survey is the product ^D2q, where ^ is the solid angle of the field of view, D is the telescope diameter, and q is the total throughput/quantum efficiency of the instrument assuming that the seeing disk is resolved by the sampling. The time required to complete a survey to a given limiting magnitude is inversely proportional to this figure of merit; the larger this 1Q product the shorter the survey time. At the time of writing the largest anticipated mosaic is the 3-gigapixel camera for the proposed Large Synoptic Survey Telescope, while the largest operational array is one of a planned set of four 1.4-gigapixel cameras for Pan-STARRS (Figure 7.14; see also Plates 15 and 16).
Construction of multi-CCD mosaics is not trivial and considerable effort has gone into both the packaging of the CCD itself and the design of the camera head in order to make these systems practical. For example, in some wide-field imagers there is sufficient curvature of the focal surface that either the CCDs must be laid out on a
Figure 7.13. The layout of eight custom CCDs packaged into a mosaic overlaid on an image obtained with this camera system. Credit: Gerry Luppino.
curved surface or a large field-flattening lens must be introduced. Even if the focal surface is flat, depth of field is another factor. For fast optical systems the depth of field is shallow and the CCDs need to be in a common plane to better than ~25 pm or one-thousandth of an inch. Every chip needs to be cooled efficiently, and there must be an easy way to remove and replace individual devices. Thus, packaging is a critical issue (Burke, Jorden, and Vu, 2005).
Was this article helpful?