LDR physics

[Martin Czech]

I had a deeper look into LDR physics. According to the text I read
the main effect in CdS LDRs (the red/orange looking types) is that
after carrier generation nearly all holes get trapped, i.e. can not
move away. This creates an amount of charge which changes the inner
energy structure of the crystals and this in turn lowers the resistance.
So the better conductivity is not so much by the few generated electrons,
but due to electrons comming in from the wire due to conditions in the
crystals. Now it is clear why LDRs turn on quite fast, but need time to
get off. The trapped holes need time to disappear due to recombination.

There is a time graph of a CdS LDR with light intensity as parameter.
A rectangular light pulse is assumed. The graph is dual log.  With 5000lx
(lux , old light intensity unit, doesn't matter here) the device needs 2ms
to saturate (curve gets flat at drop down to 10 Ohm). 500lx take 10ms,
50lx 100ms and 5lx 1000ms ! Turn on time also depends on intensity.
So there are four graphs comming down at time 0 from about 10^8 Ohm,
each with a different slope. Then, during the pulse the graphs remain
constant, each on his own level (5 min on time). When the light goes off,
each curve goes up, the interesting thing is that they all join at about
1s into a single curve.

For 5000,500 and 50lx the curves are running pretty parallel in the
beginning, only the 5lx curve is flatter. 


For static conditions the following equation can be found:

R2/R1=(E1/E2)^gamma

where E1 and E2 are the light intensities (should be linear dependend
on LED current).

gamma is the light dependency, it depends on the excact powder recepy
(CdS is usually brought to the sensor surface as powder of very little
crystals, it is not monolithic).

The larger gamma, the more sensitive the sensor will be.

I have tryed to match such a curve to a real CdS LDR, using the LED
current for E1 and E2, but the matching was not so good.

So far I have found nothing about temperature.  Temperature will of
course also generate electron hole pairs, like in any semiconductor,
but it will also affect trapping. I gues the resistance has negative
tempco.

m.c.