How to detect a single photon. How a photo multiplier tube works.

[BrianHG]

Man, a single photon.  Even in the blackest black, there has got to be some occasional detection...

[Andy Chee]

Now you know how neutrino scientists feel!

[ebastler]

Man, a single photon.  Even in the blackest black, there has got to be some occasional detection...

There's even the occasional detection without any photon coming in. 

"Dark counts" due to spontaneous electron emission from the photo-cathode are an issue and limit the ability to detect very weak signals. Strongly cooling the PMT helps, but makes for a more akward setup, having to avoid condensation on exposed cold parts.

[Phil1977]

And what´s really crazy is that this in reality is a "in average each 4th photon detector" - the overall quantum efficiency rarely exceeds 25% *.



*For the experiments these devices are used for it doesn't matter. It´s important to have a stable timing correlation to single photons, not to detect each photon.

[BrianHG]

Still, that green laser diode, not even lasing, barely visible under dark conditions by a camera, at distance, going through that pin-hole opening was clearly detected.

Since there are random empty junk detections, I wonder if you used 2 tubes with an and-gate to eliminate noise and a beam splitter at the input to drive the 2 tubes in parallel.  With the proper spacing, maybe only entangled photon pairs will be detected.

[Someone]

Since there are random empty junk detections, I wonder if you used 2 tubes with an and-gate to eliminate noise and a beam splitter at the input to drive the 2 tubes in parallel.  With the proper spacing, maybe only entangled photon pairs will be detected.

Well studied: https://en.wikipedia.org/wiki/Higher_order_coherence#Hanbury_Brown_and_Twiss_experiment

[electr_peter]

Excellent video
Scope sample rate drops at slow timebases which results in missed events or events with varying amplitude.Peak detect sampling mode would be better in this case.

I have always wondered - if input light (photon) intensity is high, do output impulses start to overlap? Or are there some pulse-relax-pulse behavior with these photomultiplier tubes?

[ebastler]

I always wondered - if input light (photon) intensity is high, do output impulses start to overlap? Or are the some pulse-relax-pulse behavior with these photomultiplier tubes?

There is a dead time while the PMT recovers (recharges its dynodes) before it can register the next photon. So the counts/intensity ratio count rate gradually saturates.

If you go much higher with the incoming light intensity you may longer see pulses at all, since instead of individual electron avalanches you get a continuous flow of electrons, with a current proportional to the incoming intensity. (Until that saturates as well eventually.)

It is not uncommon to have readout electronics which are switchable between photon counting & analog measurement of photocurrent to get a larger dynamic range.

EDIT: Sloppy wording fixed.