CHARGE-COUPLED IMAGE SENSOR

The charge-coupled image sensor (CCIS) can also be called a charge-transfer image sensor (CTIS). It is a specialized photodetector version of the parent charge-coupled device (CCD) or charge-transfer device (CTD), and is sometimes simply referred by these generic names. Due to its structure, the CCIS is also called an MOS (metal-oxide-semiconductor) photodetector or an MIS (metal-insulator-semiconductor) photodetector. When the concept of CCD was introduced by Boyle and Smith in 1970, the possibility of using it as an imaging device was briefly mentioned in their seminal paper. This idea sparked intensive research activities immediately in both CCD and CCIS. The CCIS as a linear scanning system was first demonstrated by Tompsett et al. in 1970. This was later extended to an area scanning system, presented by Bertram et al. in 1972. To extend the wavelength beyond that detectable by Si devices, compound semiconductors started to be examined in 1973. Since the 1970s, the CCIS has developed into a mature technology for two-dimensional imaging systems such as those used in cameras.

The structures of the CCIS are similar to the CCD, with the exception that the gates are semitransparent to let light passing through. A common material for the gates is polysilicon. Alternatively, the CCIS can be illuminated from the back of the substrate to avoid light absorption by the gate. In this configuration, the semiconductor has to be thinned down so that most of the light can be absorbed within the depletion region at the top surface, and that spatial resolution is not lost since each pixel typically is only about 10 nm on the side. Unlike other photodetectors, CCISs are usually closely spaced to one another in a chain. This is due to the unique feature that besides being photodetectors, they can function as shift registers like regular CCDs. This thin layer (a 0.2-0.3 nm) is fully depleted and the photo-generated charge is kept away from the surface. This structure has the advantages of higher transfer efficiency and lower dark current, from reduced surface recombination. The penalty is smaller charge capacity, by a factor of 2-3 compared to the surface-channel type. The most common semiconductor used for CCISs is Si. The insulator for Si devices is usually thermally grown oxide, and for compound semiconductors they are deposited films. Examples are ZnS for HgCdTe devices and deposited SiOj for InSb devices. The gates can be made of metal, polysilicon, or silicide. The CCIS is a unique photodetector in that there is no external DC photocurrent during light exposure. The photo-generated carriers are integrated during light exposure, and the signal is stored in the form of a charge, to be detected later. This is somewhat similar to a photodiode (p-i-n or Schottky) operated under an open-circuit condition. Since each CCIS is basically an MIS capacitor, it has to be operated in a non-equilibrium condition under a large gate pulse. If the semiconductor is allowed to recover from deep depletion, the collection of the photo-generated charge would not be efficient.