"However, Johnson (resistor) noise in the following preamplifier dominates the noise out of PIN photodiode detectors by perhaps 100 times, so any avalanche gain initially improves performance no matter how noisy the APD. Silicon APDs generally have much lower noise than III-V APDs and deliver very much more usable gain. "
Really?
It's silly to quote a number, but amplifier noise will dominate photocurrent noise at very low values of primary photocurrent. For instance, the Hobbs ElectroOptics QL01B is shot-noise limited above 7 nA in a 1 MHz bandwidth, and Johnson noise limited below that.
A multiplication gain M increases the signal without increasing the noise much, so the SNR initially goes up by 20*log(M) dB.
I recently did a special QL01 model for a large customer using a small APD, which is shot noise limited above about 5 pA in 200 kHz, running at
-10C. It's intended for PMT replacement in analytical gear. (PMTs are amazing, but not particularly cheap.)
And silicon is a much better avalanche multiplier than InGaAs or other three-fives because the ionization cross-section for electrons is about
100x larger than for holes, vs. about 10x for three-fives.The useful avalanche gain is set by the runaway increase in the noise due to the hole avalanches making electrons, which make holes, so the avalanche just bounces back and forth without dying out rapidly. (That also limits the speed.)
Some silicon APDs running at very low current continue to improve up to a gain of 100, whereas InGaAs ones top out at around 10. Still pretty useful sometimes.
Cheers
Phil Hobbs
Have something to add? Share your thoughts — no account required.
Ask the community — no account required