For many years, CCDs had much higher quantum efficiency in the Red than in the blue because the devices were physically thick (200-300 microns of material). In the blue, the absorption depth for photons in silicon is only about 2000 angstroms (1/2 their wavelength) and typically the electrodes are insulated by poly-silicate gates which are around 5000 angstroms thick. Thus, in the front-side illuminated CCD, many blue photons are absorbed by the gate material, prior to them reaching the electrode. Front-side illuminate CCDs were commonly used by observational astronomers throughout the 80's. It was quickly realized of course, that if it were possible to thin the CCD from its original thickness of 300 microns down to around 10 microns, and expose the backside of the CCD to blue photons, then those photons would reach the electrode without being absorbed by the gates. This would greatly increase the blue and ultraviolet, QE of the device without sacrificing much QE in the red (if you make the CCD too thin, then long wavelength photons won't interact with the silicon at all). Unfortunately, to produce a useful thinned CCD required that the thinning process be uniform. Variations in the thickness of a thinned CCD produce substantial focus variations across the detector. This "potato chip" factor significantly limited the performance of thinned CCDs until about 1992 when the thinning process became perfected.

The final set of improvements in CCDs is directly related to significant gains which have been made in the silicon processing industry. For highly technical reasons, it is desirable to dope silicon with ions in order to suppress surface effects related to charge transfer and the development of dark current. In addition, for back-illuminated devices, this ion implantation helps to insure that the backside of the CCD is negatively charged, thus driving signal electrons directly towards the front surface. Ion implantation techniques have been refined to the point that precision boron doping is possible. The end result is todays inverted phase CCD (this means that the substrate electric potential is larger than that of the surface) which has very low dark current even though it operates at temperatures substantially above the 77K temperature of liquid nitrogen.

B. Wide Field Imaging

Despite the improvements mentioned above, the pixel size of CCD detectors is still quite small and hence, on almost any optical platform, the field of view will be quite limited. A case in point is the Lynxx CCD from SpectraSource instruments. In its 194x165 format with 15 micron pixels, the field size through an f/10 8-inch telescope is only 4.94 x 4.18 arcminutes. Since the pointing of a typical amateur telescope is not this accurate, CCD observing can be rather inefficient as much time and effort is spent on target acquisition. The professional astronomer faces a similar dilemma in that most nearby galaxies, which have been used to define what we think we know about galaxy structure and evolution, are way too large to fit in a single CCD exposure. So how can the professional or the amateur achieve a larger field of view? Well, there are basically two choices;

The first alternative is very expensive while the second alternative requires creativity. In what follows, we offer several examples and images taken with wide field CCD systems.

Page 3


The Electronic Universe Project
e-mail: nuts@moo.uoregon.edu