On a sunny day (Sun, 9 Oct 2011 14:06:16 -0700 (PDT)) it happened Bill Sloman wrote in :
Also remeber that the PMT glass itself is a scintillator, so that will produce burst too. No EM required.
I see.
Do you think the mechanism causing the signal granularity is different in PMTs from microcalorimeters, then? Namely the same phenomenon happens there. You shine light on an absorber and measure its instantaneous temperature rise. At decreasing illumination level the observed temperature rise becomes granular, i.e. occurs in discrete jumps.
When you shine in 1.55um laser light, dim enough, you always observe
0.8 kelvin temperature jumps, which become more sparse when light intensity is reduced. When you shine in 670nm laser light you observe 1.85 kelvin jumps, with 830nm laser you observe 1.49 kelvin jumps, and so on.(Actually it is more complicated than that because laser light is in in coherent state rather than number-squeezed state. I tried to keep the argument simple, and I agree that you can argue I'm covering up some loophole with my simplification.)
What about STJ detectors? There shining in 670nm laser light seems always to produce the signal current in bursts of 800-1000 charge carriers. 830nm laser light always produces bursts of 700-800 charge carriers per burst.
(Again cutting some corners)
I agree, that regardless of the different physical mechanism being at play with each detector type, one can claim that the EM wave is in each case strong enough to knock some virtual electron off some analog of the photocathode. Then there is some unspecified internal amplification mechanism that produces, say, 800-1000 charge carriers of the STJ case out of the initial single electron. This works, *provided that* the size of the initial virtual photoelectron depends on the illuminating wavelength. Indeed, the energy of the virtual photoelectron must be hbar-omega.
Of course it is also possible to say ejection from photocathode is at play in PMTs, and other detector types are Not My Department.
Regards, Mikko
On a sunny day (Mon, 10 Oct 2011 02:39:46 -0700 (PDT)) it happened Bill Sloman wrote in :
gamma spectrometer, a diode in 'geiger' mode cannot see different amplitudes.
On a sunny day (Mon, 10 Oct 2011 15:39:42 +0300 (EEST)) it happened Okkim Atnarivik wrote in :
Interesting, never played with one, but 'temperature' is related to motion of atoms, is related to their interaction with each other, is related to the electrostatic coupling (valence bands) of the electrons, so why should that surprise me?
Well Planck's constant if found in many of those experiments. It is all EM *waves* interacting with matter, to generalise the issue a bit. That does not conflict with photon not existing, it (photon) is just a mathematical concept. You cannot really think of EM waves as one of more 'photons', that makes no sense. You will end up with a lot of nonsense. Not even worth wasting time on. But you can sure do math with p = h.v. You can also do math with fairies, 1 + 1 = 2. that does not make those real, or Elvis if you like,
I've done gamma ray detection using a diode. You can in fact get _some_ knowledge of energy. I was using a Burr Brown ACF2101 at the time to integrate and measure charge in each pulse. Charge tells you something.
Jon
, t
it.
thematical
no sense.
Specific experiments do illustrate both side of the wave-partical duality.
If you try to keep both in mind at once, you will end up confused, which is where you seem to be.
You can't waste a lot more time by trying to ignore the duality.
Both correct. But they don't lead to your conclusion.
-- Bill Sloman, Nijmegen
Sloman
udes.
But not in Geiger mode - you kill the avalanche as soon as you detect it, and with it any charge information.
-- Bill Sloman, Nijmegen
Once I changed opamps in an integrating sink. Low frequency noise was all important to this design. It reduced fluctuation a bunch. Unfortunately the experiment in this lab stopped shortly after. I would not be surprised if someone is still using the old design.
Greg
Sloman
The ACF2101 holds the diode against a virtual ground and uses switchable capacitors on the feedback, with carefully balanced switches regarding charge injection and removal when switching. I placed it in an ice water bath to stabilize the detector temperature (and the circuit, as well) and took weeks of interesting data using a variety of gamma sources at various incidence angles to the detector. It was an interesting experience and I learned a few things from it. One was that penetration laterally across the diode yielded a much larger collection than perpendicular to the face.
Jon
Nah, they're a good six orders of magnitude noisier than a PMT, and they're slow as molasses--the rise time can be reasonable, but the fall time is the pits.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
My experience with cosmic rate detections on diode detectors at sea level were on the order of 0.5 to 1.5 minutes, using a
1cm x 1cm Hamamatsu detector at virtual ground. 100pF detector at sub-millivolt virtual ground, if memory serves. I forget the integration rate, now, but I recall calculating on the order of 500 to 20,000 electrons per event. Notes are somewhere in CA, now, and not accessible to me anymore.Jon
The noise level is higher, though to get six orders of magnitude difference you would need to compare a cooled blue-sensitive PMT with a germanium photo-diode.
With active recovery, they can be quite fast
talks about 40nsec dead time and a maximum counting rate of 10MHz with a "thick" diode, and 10nsec and 40MHz with a "thin" diode. IIRR his active recovery circuit included an AD96685 comparator, or something very like it.
Sergio Cova seems to make his own avalanche photodiodes
There's a bit of energy temporarily stored in the avalanche photodiode after every avalanche which does show up as an occasional after-pulse, but you've got the same kind of problem in a photocathode.
-- Bill Sloman, Nijmegen
Low frequency noise is one of those dirty sorts of problems we'd all much rather avoid. You can't filter it out, you can't null it out, you just have to measure all sorts of small effects and pick the right parts. Even then, you have to temperature-control stuff carefully, because at low enough frequency, temperature drift usually dominates.
On favourite part of mine is the OPA378--a sort of CAZ-chopper hybrid that has about 30 nV noise right down to DC. That's better than even a quiet bipolar, at least down in the millihertz and below.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
l
Except that you buy devices, not detector area.
Perhaps, but in practice you don't buy area, you buy devices. Large area detectors are weird, exceptional, expesnive and you usually have to wait until the manufacturer can make some for you.
If you need the large detector area. Many applications don't. You could get a magnets to stick on the photocathode of a conventional 2" dimaeter tube so that only the electrons from the central bit of the photocathode made it to the first dynode.
Fairly obviously, since the paper I was pointing to was published in
1996.
If you want a large sensitive area, which isn't all that common.
10MHz is actually pretty fast for an ordinary venetian-blind PMT - there's a lot of stray inductance in the leads of a regular PMT, and the whole structure looks like a bunch of resonant circuits if you try to run them fast.Micro-channel plate multipliers were pretty exotic when we used them at Cambridge Instruments in the 1980's - they solved a number of problems in the electron-beam microfabricators, but you needed a new one every six months, which sort of worked for a machine that sold for a couple of million dollars and went into a semiconductor fabrication line. I can't recall every seeing one in a regular photomultiplier catalogue.
..
Why this enthusiasm for sensitive area? Most applications seemed to stick the detector at a focal point.
Good physicists do pay attention to solid angles and active areas, but engineers are usually happy with something that works.
-- Bill Sloman, Nijmegen
Not me, and not any other competent designer either. And that factor of a million--a MILLION, now, not ten percent--means that even if you can't find a PMT as small as you'd like, it'll nearly always be a big SNR win.
Do you know of anyone who stocks a boatload of different APDs so you can always find just what you want? It's all a niche business.
Competent instrument designers don't buy the wrong thing, especially when the cost and performance deltas are so very large.
In your world, I imagine not. Ultrasensitive measurements are what PMTs are for.
It isn't the whole structure you have to worry about. Regardless of the Q, you aren't going to make a significant change in a 100 volt dynode-dynode bias with a current of a microamp. 30 MHz is routine for photon counting with a PMT.
Depends on the application. They do have much shorter lifetimes than PMTs measured in anode coulombs, as I mentioned. but in very low level measurements they can last a long time.
If the application is that easy--most aren't IME. But that factor of a million means that a 100 micron diameter APD has the same noise as a four-inch PMT. Use a 6 mm tube instead of a 100-um diode, and you get painless alignment and still save a factor of almost 300 in the dark count rate.
If you think that eating a factor of 10**6 for no reason is good engineering, I have news for you.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
te:
eall
nNonsense. You don't go out and commission the design of a new detector for every new application, you design around what you can buy.
If you need it.
1That's what I've just been saying.
lid
eDs,
Fad-addicted instrument designers make a habit of it.
hng
a sUltra-sensitive to what? At Cambridge Instruments we were a relatively large customer for photomultiplier tubes because each scanning electron microscope contained one.
It would have been nice to collect the scintillator output over a larger solid angle, but a scanning electron microscope selling for about $100,000 at the rate of about a hundred a year wouldn't justify a special-purpose part.
d d eAnd how fast can you actually count before your "single photons" overlap to an embarassing extent?
In a good vacuum.
5.... e ,he
If it matters. In many applications you've got enough signal that the dark count rate is less than the shot noise, and doesn't matter.
And if you think that that the factor of 10^6 matters in every applications, I have news for you.
-- Bill Sloman, Nijmegen
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.