Can a Photomultiplier be Calibrated for Extremely Low Light Levels?

[MereA.]

Does anyone know how to calibrate a photomultiplier? I need to measure very, very low levels of light (almost no light) and am considering using a photomultiplier. However, I don't know how easy (or difficult) it would be to calibrate a photomultiplier to be sensitive to such a low level of light. If a photomultiplier won't work, are there other options? I've looked at lux meters but could not find any that would measure below 10^-2 lux, which is too much light. Thank you for any help.

 

[NobodySpecial]

Typically an Fe55 source which gives an x-ray photon of very consistent energy which will generate a known number of electrons in the photo cathode.

An simpler method is a pulsed LED source and a stack of ND filters

 

[MereA.]

Do you know if there is a good source (website, book) that would describe the process of calibration using an LED source and ND filters?

 

[Dickfore]

In what physical quantity do you want it calibrated?

 

[MereA.]

If possible, I'd like to express the amount of light in terms of lux.

 

[Dickfore]

You do know that the lux is a photometric unit of light intensity, right? As such, it is only valid for visible light and you have to find the SI definition of candela (through one standard source that I do not remember from the top of my head).

 

[NobodySpecial]

Light is annoyingly difficult to calibrate accurately - better than the few % level.

Easiest way to do it is to compare a source with a light meter then use a stack of photographic ND filters to drop the source by a known amount for your low light sensor.

 

[MereA.]

Okay. Thank you for your help, NobodySpecial and Dickfore.

 

[Bob S]

²This url has a table that converts lux to photons.

http://www.egc.com/useful_info_lighting.php

For example, from the table, 54 sunlight photons (per m² per sec?*) is equal to 1 lux. It is possible to calibrate the gain of a photomultiplier tube (PMT) by putting in a dark box, and by using a pulsed light source and neutral density filters (NDFs), get the light level so low that the distribution of number of photoelectrons from the PMT cathode is given by the Poisson distribution (i.e., less than 10 photoelectrons per light pulse) with discrete pulse heights. You will need a gated pulse height analyzer for this. By attenuating the light another factor of 10 using a NDF, nearly every output pulse is due to a single photoelectron. measuring the Coulombs per photoelectron is straight forward. For example, a single photoelecton in a 10-stage PMT might create about 10^-13 Coulombs at the anode.

* Something is very fishy. 1 watt of sunlight per square meter is about 2.5 x 10¹⁸ photons per square meter, from basic physics.

[added] the table in the url is in mcromoles of photons per m² per sec, or 6 x 10¹⁷ photons per m² per sec. Using the conversion to watts in the table gives 4.57 x 6 x 10¹⁷ = 2.74 x 10¹⁸, in agreement with the above physics example.

Bob S

 

[Dr Lots-o'watts]

...If a photomultiplier won't work, are there other options?...

These devices can compete:
http://en.wikipedia.org/wiki/Avalanche_photodiode

You may also want to browse in our own selection of instruments, just in case.
www.gentec-eo.com (from fW to kW)

 

[NobodySpecial]

APDs are generally easier than photo-multipliers if you don't need a very large area. They don't need scary HV supplies, have better red sensitivity and are pretty immune to damage from high light levels

The electronics for them is pretty simple - I had breadboard circuits doing individual photon counting with avalanche mode APDs 20years ago.

 

[Bob S]

Here is a good selection of Hamamatsu photomultipliers:

http://sales.hamamatsu.com/en/produ...s/photomultiplier-tubes/productlist.php?#form

As one example, the Hamamatsu R760 tube has a gain of about 1 x 10⁶, and a dark current of about 1 nA, meaning the dark current is equivalent to about [STRIKE]600[/STRIKE] 6000 cathode photoelectrons per second. If you could gate the signal from the PMT, you could get a clean signal without a dark current background.

Bob S

 

[Vanadium 50]

PMTs are terrible at measuring absolute light levels, particularly high gain tubes. Truly horrible. First, you are limited by the statistics of photoelectron yield at the first dynode, and second, the quantum efficiency of your photocathode varies enormously from tube to tube and even across the face of a single tube. Calibrating on single photons and scaling up introduces its own uncertainties - you have to worry about dark current and thermionic emission (i.e. current from non-light) and emission off the second dynode (i.e. reduced gain).

The only time people use them is when they have no choice. If you go that route, you will be spending 95% of your time understanding your instrumentation and 5% of your time doing your measurement.

One question you should have been asked - "How much light are you measuring?" The other is "How accurately do you need to measure it?"