> From: "Robert A. LaBudde"
...
> an absolute measure of radiant light power at even a single fixed wavelength
...
>how does NIST calibrate the meter in the first place?

I see that Stephen R Phillips, Mariano Alvira, "tony@spiderco", and
Harold Hallikainen all gave excellent, practical information on how
this is is actually done in the real world
(in general, absorb the light with some black material, and measure
the temperature rise).

But that's not going to stop me from rambling on about a half-baked
theory on another possible way to do it. (I am not a physicist, so if
there is some reason this couldn't possibly work, please point out the
fatal flaw).

Starting with:
* an (uncalibrated) reference light source that gives out relatively
monochromatic and relatively constant light power. (Perhaps a narrow
laser beam).
* a spectrometer to measure the frequency of the light source, and to
confirm that it is close enough to monochromatic
* an (uncalibrated, and possibly nonlinear) light power meter.
* a photodetector with a known quantum efficiency (for example., some
high-quality CCDs have a quantum efficiency of 9 output electrons for
every 10 photons that pass through the aperture).
* some half-silvered mirrors that each split the input light close
enough to 50:50 with negligible loss. (I think it's possible to
confirm "negligible loss" and "close to 50:50" even with an
uncalibrated, nonlinear light power meter).

You then
* set up a chain of half-silvered mirrors with output ports that emit
1/2, 1/4, 1/8, 1/16, 1/32, etc. the input light. (a bunch of output
laser beams, each one progressively weaker than the one before).
* Put the photodetector far down the chain so you can count individual
photons at a reasonable rate (e.g., if it's "clicking" at 900
clicks/second, and you have a quantum efficiency of 0.900, then you
are getting individual photons at 1000 photons/second).
* put the light power meter closer to the start of the chain so it's
at high enough intensity to make accurate measurements
* calculate the rate that photons hit the light power meter -- (rate
that photons hit the photodetector) * 2^( number of steps in the chain
between the photodetector and the light power meter).
* calculate the physical power hitting the light power meter -- use
the photon frequency you measured with the spectrometer, the Planck
constant to convert that frequency to energy per photon, and multiply
by the rate that photons hit the light power meter.
* mark the current location on the power meter's output dial with the
physical power you calculated.
* move the power meter up and down the chain and mark its output dial
with the other known beam powers ... this lets you see if the power
meter indicator dial is linear with power ... and if it is not linear,
find out what the curve is and (hopefully) compensate for it.

Alas, I think this procedure merely takes 1 question and makes 2 more out of it:
* How is the Planck constant measured accurately?
* How is the quantum efficiency of a particular photodetector at a
particular wavelength measured accurately?

--
David Cary
1.918.813.2279
http://carybros.com/
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist