Generation recombination (GR) noise

I've been working with PEM sensors

which are fast (nanosecond) detectors that work at thermal/CO2 wavelengths, like 10 microns. The mechanism is bizarre, essentially photon induced Hall effect. They generate a few microvolts just from the IR of my hand nearby.

I have successfully evaded a play about a crazy woman, a 3-hour opera, and a ukelele recital.

I think we used those with a CO2 laser.

I wonder if a CdS photocell shows basically GR noise?

Well both "my girls" (Wife and daughter) will be dancing. I don't mind that. But there will also be hours of ~six year olds in tutus, which is cute for about five minutes.

George H.

I believe the argument goes like this:

The shot noise is fast, but the recombination is slow. Any positive-going fluctuation in the mean carrier density obviously dies out just the same as a signal fluctuation. Since in a steady state recombination balances generation, negative fluctuations behave the same way, or else the mean photocurrent would walk continuously upwards.

This part of the noise gets multiplied by the photoconductive gain

G = lifetime/transit time .

There's also a fast component of the shot noise, because the instantaneous photocurrent responds much faster than 1/lifetime, but that part doesn't get multiplied.

So you have a two-plateau noise spectrum, basically

n(f) ~ sqrt(4 e I_photo) * (1 + G/(omega*lifetime))

where the 4 in the first term comes from the generation and recombination both being stochastic, unlike the case in photodiodes, where recombination occurs deterministically when the carriers reach the contacts.

Essentially this is a gain-bandwidth tradeoff of the sort we're used to, with no free lunches, as one would expect.

Cheers

Phil Hobbs

Oh, and of course you have to add the white thermal noise in RMS, which makes the formula messier.

Cheers

Phil Hobbs

Gamma ray detection is nicer. BANG. No doubt about it. It's all shot noise.

I guess the shot noise is lower for the super long wavelength detectors, especially since their sensitivity is low.

Thanks Phil, more thought is required. (I'm going to save it for tonight in the dark auditorium, as a sea of tutus flutter about.)

George H.

Little girls are cute, in reasonable quantity.

Yes, provided "reasonable" is defined as "fewer than 4, not counting our own offspring".

School concerts are a refined form of torture. They're horrible, but (a) you have to pretend to like them, and (b) the teachers use that against you for next year.

Cheers

Phil "my youngest is 23 now" Hobbs

I'm lucky in that both of my girls were tomboys who couldn't be bribed into wearing a dress. It was all about horses and baseball bats. With girls like that, who needs boys?

I have two daughters and one son. They're very different.

Cheers

Phil Hobbs

The chorus sounds nice from the start, the band is OK after ~7th grade, and can rock in HS the orchestra only gets good in HS.

There is a end of year All Choral Concert, which I look forward to. They have the choruses from the grade schools (2) middle school and high school all there. They each sing a few songs, (sometimes mixing and matching.) And end with everyone singing together... always the same two songs. One song is about lighting a candle.. and then the Irish Blessing. It's very touching, I've watched my little girl move up through the ranks. (wipes tear from corner of eye.)

My son plays in the band, they have a similar tradition. There is always a HS band concert before Xmas and for the last song they always play "Sleigh Ride" Any musicians in the audience (mostly alums) can bring their instrument, and join the band on stage... It's great! (put's a smile on my face.)

George H.

Hi Phil, Again thanks for that. I assume the above is for semiconductors that have some constant carrier concentration. This case is a bit different. (Which I think makes it easier... analysis wise.) All the carriers are created by FIR photons. The "background" resistance (number of carriers) is set by how much black body radiation is allowed to leak down from the top. There are several layers of black poly. (magic black plastic) That sets the back ground level. You can choose more or less.

There is then the synchronous light from the FTIR spectrometer. This is a few percent of the back ground light. So all the carriers are "created equally". Anyway I'm working out the details. (And no don't do it for me.. I'll write again if I have confusion.)

George H.

The dark current sets the Johnson noise floor. In most FIR photoconductors it's pretty large. (If you could make good diodes out there, you would.)

The funny spectrum comes from the photoconductive gain mechanism, so I think it's independent of the dark current.

Cheers

Phil Hobbs