Photdiode optimized for photovoltaic/photoconductive mode?

[Guy Macon]

I am an old hand at interfacing to photodiodes, but there is one
thing that I haven't figured out. On occasion I read about a
photodiode being "optimized" for either photoconductive mode
(reverse bias) or photovoltaic mode (no bias, photodiode feeds
current into a virtual short). What, exactly, are the parameters
that differ between the tow optimizations? What are the process
differences? It has been my experience that you can run any
photodiode in either mode and get pretty much the same results.
Perhaps this is because I usually am on the low light, large
area, low speed end and the differences are more apparent at
the high speed end.

 

[Ken Smith]

I am an old hand at interfacing to photodiodes, but there is one
thing that I haven't figured out. On occasion I read about a
photodiode being "optimized" for either photoconductive mode
(reverse bias) or photovoltaic mode (no bias, photodiode feeds
current into a virtual short). What, exactly, are the parameters
that differ between the tow optimizations?

The main difference in application is the breakdown voltage. A typical
large area detector intended for PV mode will only have a breakdown
voltage of 3 or 4V. They can still be used in photoconductive mode but
you need to use a fairly low voltage or the leakage current will be too
high.

I think that the process difference is that the PV devices are more
heavily doped, but I'm probably wrong. Both types are PIN diodes. A
section in the center is left very lightly doped.

 

[Phil Hobbs]

The main difference in application is the breakdown voltage. A typical
large area detector intended for PV mode will only have a breakdown
voltage of 3 or 4V. They can still be used in photoconductive mode but
you need to use a fairly low voltage or the leakage current will be too
high.

I think that the process difference is that the PV devices are more
heavily doped, but I'm probably wrong. Both types are PIN diodes. A
section in the center is left very lightly doped.

Two other things that can enter are dc leakage and capacitance. Lots of PV
diodes are actually more like PN types, as you say. This also means that
they have huge capacitances and don't work as well in the IR, because a lot
of the light gets absorbed down in the substrate where there's no E field to
separate the carriers. A decent PIN diode can be fully depleted, leading to
a capacitance reduction by a factor of up to about 6 or 7, which helps a lot
if you need any sort of speed or good noise performance. For that, you need
a high breakdown voltage and low leakage.

Cheers,

Phil Hobbs

IBM T. J. Watson Research Center
Yorktown Heights NY 10598