A CCD is a silicon based semiconductor arranged as an array of photosensitive elements, each one of which generates photoelectrons and stores them as a small 'bucket' of charge. Each pixel is typically 15 to 30 m square. Our first CCDs had formats of about 400 rows by 600 columns, with an area of about 1 square centimetre. Current CCDs have formats of about 1024 1024 pixels, with a size of about 25mm square. We anticipate introducing 2048 2048 pixel CCDs before the end of 1995.
When requested, the elements form a bucket brigade; each row of charges is passed from element to element, a process which is known as clocking, down the columns and horizontally along the final row. The value in each pixel is measured in turn and recorded digitally. To ensure that only positive numbers result from this analog to digital conversion process a fixed offset known as the bias level is introduced. The charge-transfer process is essentially noise-free and almost all of the noise contributed to the signal by the CCD is from the output stage, where the charge content of each bucket is measured. This is called the readout noise.
The dark current of CCDs means that they must be cooled to cryogenic temperatures for use as astronomical detectors. At room temperatures, the dark signal is such that most CCDs would saturate in a few seconds. Cooling via temperature-controlled liquid-nitrogen dewars (or by closed-cycle refrigeration) provides a crucial reduction in dark current. The mobility of electrons is somewhat impaired by cooling, so that a compromise is required to maintain adequate charge-transfer efficiency. The optimum temperature is about 150 K, depending on specific device type.