Just buy an old photomultiplier,connect it to a
sensitive (audio) amplifier, cover it up .
We did it with several layers of special black cloth.
Switch on, and you get separate pops for each leaking
photon and emitted electron. Lift a corner of the
covering slightly, and the photomultiplier starts
emitting a staccato of plops.
One plop for each electron leaving the photosensitive
layer.
OK, but it's really demoing the quantization of photons, sort of, and
it doesn't quantify the charge at all, or not unless you estimate the
dynode gains. I was thinking of actually measuring single electrons.
If you observe the output of a high-gain cmos opamp connected to a
capacitor (which could be its own internals) there will be a step for
every electron of leakage, but it will be way below the noise floor.
Since the electron leakage is random and the noise is random, I can't
think of an obvious way to signal-average the output and reveal the
steps.
If the electron leakage were triggerable, one could measure the signal
before and after each shot, and dig the steps out of the noise. Maybe
you could inject a shot of charge capacitively, statistically roughly
one or two electron's worth, and average out the resulting signal
change, and show that it's quantized into steps.
An eprom is still interesting, since the geometry and charge are so
small.
John