1.4 Sensitivity

Since a biased MOS structure builds up a depletion region with an electric field similar to that in a p–n junction, it's worth thinking about what would happen if light were to be absorbed in an MOS capacitor, generating electron–hole pairs.

The generation rate of charge carriers follows the standard in terms of the incident optical power, the wavelength and the efficiency of conversion.

Whereas photoconductors and photodiodes generate current in proportion to incident light, the MOS structure automatically integrates photocurrent into accumulated charge. The total accumulated charge after a shutter time is:

The sensitivity in this case, measured in coulombs (C) per watt, is .

Absorption of photons generates a current in a biased MOS structure; see Figure 5. Photocurrent builds up charge in the electron accumulation region in proportion to the number of absorbed photons. Here, then, is another way of translating light intensity into electronic information. Provided it is not ‘over-exposed’, an array of MOS photocapacitors can be charged in proportion to the light intensity to which it is exposed. However, the thermal generation that was discussed in section 1.3 remains active, so the two mechanisms proceed in parallel. To be an effective pixel, the photocurrent must significantly dominate the thermally generated dark current. (To distinguish them from p–n photodiodes, I have chosen to refer to MOS-based light detectors as photocapacitors.)

Figure 5 Generation of photocurrent in a biased MOS structure (note that details of the thermal generation and back diffusion of holes are not shown)