Unfortunately, the spectral responses of these devices are quite varied, depending on details of manufacturing (e.g., polysilicon vs. aluminum electrodes; thinned vs. non-thinned), preparation (use of blue-enhancing treatments such as phosphors, UV-irradiation, and flashgates), and use (front-side vs. back-side illumination). The response also depends on temperature, as it does for all semiconducting materials.
The preferred method of handling this problem is to use a table of (averaged)
spectral responsivity determined for the individual device under actual
conditions of use.
This means using the spectral-response table file described under
DET= OTHER (section 
).
The name of the table file should be contained in a character column
named
SPECTRESPTBL, of adequate width to hold a file name (see
section for a description of the file).
If no table is available, this column can be omitted.
If tables are available for only some detectors, the column should exist, and
contain blanks in the rows of detectors lacking detailed information.
Even if no detailed spectral response is available, it is sometimes necessary to make a rough guess at the spectral response, particularly for planning purposes. Then, if SPECTRESPTBL is missing, there should be a character column named BLUERESP that gives some indication of the expected blue response. Valid string values are 'FRONT' for front-side illumination of a normal CCD, with no phosphor coating; 'BACK' for thinned, un-enhanced CCDs illuminated from the substrate side; 'ENHANCED' for CCDs treated by UV irradiation, special gases, or flashgate; or 'PHOSPHOR' for CCDs coated with a blue-sensitive phosphor. These general categories provide some information on spectral response, but it may easily be in error by a factor of 3 or more.
As for PMTs, the COOLING column is required. Usually, CCDs are cooled with liquid nitrogen, but are actually kept thermostatted by a servo system. In this case, COOLING= 'REGULATED'. The same applies to chips cooled thermoelectrically, in most cases. Occasionally one finds unregulated cooling; then COOLING= 'UNREGULATED', and the treatment is the same as for cooled but unregulated PMTs (see above).