The P region is thin because its length must comparable with diffusion length of minority carriers injected in this region and so to avoid the recombination them. Therefor yes, it is very relevant for the charge transport.
If the PIN diode is a photodiode, and the light is incident on the P side, you would like the P side to be thin so that more light is absorbed directly in the I region, where there is a high electric field and the photo-generated electron-hole pairs contribute immediately to the external photocurrent. If the light is absorbed in the P layer, the minority carrier photogenerated electrons must first diffuse to the I region, a relatively slow process that could also have low collection efficiency (low quantum efficiency) if the distance the carriers have to diffuse in the P material is a significant fraction of the diffusion length in the P material.
Steven Wojtczuk's answer is correct, but applies (as he indicates) to photodiodes only. I would add that the diffusion length varies approximately as the square of the resistivity, so a low resistivity front surface layer (as required for fast electrical time constant and thermal noise addition) will be quite lossy unless the layer is thin. 100-ohm/sq would represent about 3-Ohms series resitance and so provide an adequately short electrical time-constant. If (for example) we use a doping of 1E19 and a 1-um thick layer the response time for carriers generated near the surface would be determined by the layer thickness as welll as the carrier lifetime, and collection efficiency for carriers generated in the P layer would be somewhat below 50%; the tail would have a time-constant in the order of 100-ns. 2E20 doping and 0.1-um depth would both reduce the proportion of carriers generated within the P+ layer and the collection efficiency for these carriers; in addition we would be looking at bstantial carrier loss within the layer which would further reduce the amplitude of the tail, whose time-constant would be about 2-ns. Further increase in P doping at the same surface resistivity would reduce the amplitude of the "slow" optical response and reduce the time-constant, both being approximately proportional to the the inverse square of the doping. In addition, a thinner P-doped layer (in a mono-junction) improves the short wavelength response, making the detector more general--purpose.
Microwave PIN diodes generally have shallow doping at the top surface; this is to minimise the doping tail, which creates a series region of low-doped (but not intrinsic) material.
- Badges
- Science topic