Photon counting w/ APD's

This is a bit of a shot in the dark, but I wonder if anyone has used 'garden variety' APD (Avalanche Photodiodes) in the photon counting mode where they are biased above their breakdown voltage?

I just got a quote on some of these,

formatting link

The difference between the SAR and SARP looks mostly to be the noise and dark current. (see figs 6,7,8)

Oh the SAR500 is $60 for one the SARP500 is $150 and with TEC they are $375 and $670... that is some serious scratch.

George H.

Reply to
George Herold
Loading thread data ...

Why APD's and not MPPC(Silicon Photo Multipliers)?

--
Uwe Bonnes                bon@elektron.ikp.physik.tu-darmstadt.de 

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt 
 Click to see the full signature
Reply to
Uwe Bonnes

I need to count photons with some decent efficiency at ~800nm. longer wavelength might work too, but low dark count and high QE is what I need. I should probably get fiber coupling too. It's an entangled photon, exp. You start with a UV laser (~400nm.) with a magic x-tal to phase entangle the down converted

800 nm photons. (I need to read a lot more about it.)

George H.

I'm going to order

Reply to
George Herold

Well, "low dark count" and APDs only go together for very permissive values of "low". On a per-area basis, APDs have six orders of magnitude worse dark counts than PMTs, and that's an apples-to-apples comparison: Si APDs and visible-sensitive bialkali PMTs--a 100-um APD has about the same dark count rate as a *FOUR INCH* PMT.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

I've not used them, but I have had some minor interaction with people who are rather good with them, including Sergio Cova

formatting link

They have long list of publications. The one I liked used the latching inputs on a 9685 (fast ECL) comparator in an active latch SPAD - which was vaguely similar to something I'd done in a very different context, but I didn't get a citation.

Sloman, A.W. and Swords, M.D. "A fast and economical gated discriminator", Journal of Physics E: Scientific Instruments, 11, 521-524 (1978).

I can't pick it out of the list - it would have been published some 15 years ago, and I can find a couple of plausible candidates, but nothing really sticks out.

formatting link

--
Bill Sloman, Sydney
Reply to
bill.sloman

Right, If you know of a small pmt with quantum efficiency (QE) greater than (?) ~50% at 800 nm I could try that. :^)

I still have to understand the experimental details, but I think I'm looking for correlations between two detectors. The higher the QE the less one has to bash on the data with statistics to see a correlation. Obviously dark counts go into that mix too. I hear I can cool 'em down and reduce the dark count. I wonder if one can stick them in LN2?

Anyway the people at Laser components were friendly and helpful. I'm going to order some of the APD's to play with.

Their fab plant is in Tempe AZ. Maybe I can have Jim T. swing by and steal me a few :^)

Here's a long discussion from Kiko at Colgate. (I still have to read and digest most of it.)

formatting link
George H.

Reply to
George Herold

Right, thanks Bill. I've read some of Cova's articles. At some point in the future you can all look forward to me asking about active quenching circuits. :^)

George H.

Reply to
George Herold

You might consider using a microchannel plate, like from a night-vision gadget. Since they are area amplifiers, one could possibly be used as two (or more!) detectors.

Even a gen-1 night vision thing (they are cheap) could be used as a photon amplifier maybe.

Hamamatsu has a tiny MEMS PMT that they don't want to sell.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Hmm I don't know about those. (guessing) There is some photocathode, that emits electrons when a photon hits it. In which case it will have the same problem as PMT's. The photocathodes are not that efficient at 800nm. The APD's I'm looking at have a QE of 90%!! at

800nm (For correlations with two detectors it's the square of the QE that I care about, so 10% is the pits.)

They have some Silicon PM (MPPC) which are arrays of APD's, but the ones I know about are all in the visible... response peaks ~500 nm.

formatting link

Trolling Hamamatsu, these might be OK.

formatting link

But no data on operation above the breakdown voltage. George H.

Reply to
George Herold

Yes. And the electrons get amplified, by a microchannel plate or, in the cheap ones, just by being accelerated by a lot of voltage.

The dark emission should be like a PMT which, as Phil notes, is zillions of times better than an APD.

In which case

The night-vision gadgets usually include a near-IR illuminator, just barely visible, so it's around 800 nm. The classic sniperscopes worked with invisible IR, so the bad guys wouldn't see the immuminator.

I have a cheap russian night-vision scope that's great for spying on critters in the back yard at night. Skunks, raccoons, cats, possom, now occasionally coyotes, plus deer and wildcats in Truckee.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Right, this experiment is all about two correlated photons, down converted with a x-tal from a 405 nm laser diode. The correlation goes as the product of the detector eff. So at 10% QE only 1% of the 'good' events are detected. I guess anything above 50% should be good enough, but you always want as much as you can get... signal wise.

You can buy these detectors for ~$2-5k*, the price depends on the dark count. I've seen numbers from 10 Hz to 10 kHz. (. at some point it's a reversed biased diode above the avalanche point and it breaks down continuously... it only stops conducting at lower bias 'cause the avalanche is a little random. (or someone shuts it down.) at least that's my limited understanding.) There's after-pulsing too, charges stuck in the channel, or something.

George H.

*laser-components makes packages too, I asked for prices today... I hate when there is no price list. It's a complicated physics package.
Reply to
George Herold

That rather depends on the photocathode. Photocathodes that emit electroncs when hit by 800nm photons have pretty high dark currents compared with photocathodes optimised for shorter wavelengths.

They also tend to need to see a lot of photons before they'll kick out one electron.

What are they supposed to say - if you don't limit the current when operating above breakdown voltage, you fry the device?

Passive and active quenching does exactly that.

--
Bill Sloman, Sydney
Reply to
bill.sloman

Avalanche multiplication is a statistical process, and if you have only got one electron going through the avalanche area, it can make it all the way without creating a second charge carrier pair to sustain the process.

Avalanche breakdown diodes - sold as Zener diodes though the Zener mechanism is only active for breakdown voltage less than about seven volts - are good noise sources at low current precisely because the avalanche self-quenches from time to time.

We had a long thread about this many years ago, with Win Hill posting lots of literature references, and Tony Williams posting lots of measurements.

--
Bill Sloman, Sydney
Reply to
bill.sloman

You can do that. You use a prism to launch light into the faceplate at greater than the critical angle, so it bounces around until it gets absorbed in the photocathode. It's photocathode transparency that's the big QE hit in end-on tubes--the PC is thin enough that the quantum yield of photons that actually get absorbed is pretty good.

Side lookers have opaque photocathodes, so they're a bit better for QE but have other problems such as huge sensitivity variations with position. Also their PCs are thick, so quite a lot of photoelectrons don't make it to the surface.

Leaked excerpt from BEOS III regarding prism coupling: ========== This method can reach QEs of at least 58\%, comparable to uncoated photodiodes.\footnote{W. D. Gunter Jr., G. R. Grant, and S. A. Shaw, "Optical Devices to Increase Photocathode Quantum Efficiency", \emph{Appl. Optics} {\bf 9}, 2, p. 251 (1970)} A simpler scheme is oblique incidence plus a return mirror, which at least gets you two passes. Gunter \etal\ also show an interesting combination: a solid cone concentrator can redirect most of the incident light into TIR rays without reducing the photocathode area.

The prism method is quite a bit better, because the transmitted wave is totally reflected at both the glass/air and photocathode/vacuum interfaces, so it has to rattle round until it's detected, and (most interestingly) you avoid the transmission loss as well as the reflection loss at the photocathode---{}you get two oblique passes through the photocathode per bounce. Photocathodes, especially NEA ones, are pretty good at generating photoelectrons from photons they actually absorb, so you can get factors of 2 to 5 in quantum efficiency this way, depending on the wavelength. Given how much you're paying for those photons, this is usually a very good deal. =========

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

That's an interesting idea, Thanks Phil... I know you've mentioned it before, but I never quite understood. One potential problem I see for my application is the photon rattling around in the glass interface, is going to add some unknown time delay. That may be a non-issue,

100ps probably won't matter.

George H.

Reply to
George Herold

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.